Articles

Optimal fractionation of preoperative radiotherapy and timing

to surgery for rectal cancer (Stockholm III): a multicentre,
randomised, non-blinded, phase 3, non-inferiority trial
Johan Erlandsson, Torbjörn Holm, David Pettersson, Åke Berglund, Björn Cedermark, Calin Radu, Hemming Johansson , Mikael Machado, Fredrik Hjern,
Olof Hallböök, Ingvar Syk, Bengt Glimelius, Anna Martling

Summary
Background Radiotherapy reduces the risk of local recurrence in rectal cancer. However, the optimal radiotherapy Lancet Oncol 2017
fractionation and interval between radiotherapy and surgery is still under debate. We aimed to study recurrence Published Online
in patients randomised between three different radiotherapy regimens with respect to fractionation and time to surgery. February 9, 2017
http://dx.doi.org/10.1016/
S1470-2045(17)30086-4
Methods In this multicentre, randomised, non-blinded, phase 3, non-inferiority trial (Stockholm III), all patients with
See Online/Comment
a biopsy-proven adenocarcinoma of the rectum, without signs of non-resectability or distant metastases, without http://dx.doi.org/10.1016/
severe cardiovascular comorbidity, and planned for an abdominal resection from 18 Swedish hospitals were eligible. S1470-2045(17)30075-X
Participants were randomly assigned with permuted blocks, stratified by participating centre, to receive either 5 × 5 Gy Department of Molecular
radiation dose with surgery within 1 week (short-course radiotherapy) or after 4–8 weeks (short-course radiotherapy Medicine and Surgery,
with delay) or 25 × 2 Gy radiation dose with surgery after 4–8 weeks (long-course radiotherapy with delay). After a Karolinska Institutet and
Centre of Digestive Diseases,
protocol amendment, randomisation could include all three treatments or just the two short-course radiotherapy Karolinska University Hospital,
treatments, per hospital preference. The primary endpoint was time to local recurrence calculated from the date of Stockholm, Sweden
randomisation to the date of local recurrence. Comparisons between treatment groups were deemed non-inferior if (J Erlandsson MD,
the upper limit of a double-sided 90% CI for the hazard ratio (HR) did not exceed 1∙7. Patients were analysed Prof T Holm PhD,
B Cedermark PhD,
according to intention to treat for all endpoints. This study is registered with ClinicalTrials.gov, number NCT00904813. Prof A Martling PhD);
Department of Molecular
Findings Between Oct 5, 1998, and Jan 31, 2013, 840 patients were recruited and randomised; 385 patients in the Medicine and Surgery,
Karolinska Institutet and
three-arm randomisation, of whom 129 patients were randomly assigned to short-course radiotherapy, 128 to short-
Department of Surgery,
course radiotherapy with delay, and 128 to long-course radiotherapy with delay, and 455 patients in the two-arm Norrtälje Hospital, Norrtälje,
randomisation, of whom 228 were randomly assigned to short-course radiotherapy and 227 to short-course Sweden (D Pettersson PhD);
radiotherapy with delay. In patients with any local recurrence, median time from date of randomisation to local Department of Immunology,
Genetics and Pathology,
recurrence in the pooled short-course radiotherapy comparison was 33∙4 months (range 18∙2–62∙2) in the short-
Uppsala University, Uppsala,
course radiotherapy group and 19∙3 months (8∙5–39∙5) in the short-course radiotherapy with delay group. Median Sweden (Å Berglund PhD ,
time to local recurrence in the long-course radiotherapy with delay group was 33∙3 months (range 17∙8–114∙3). C Radu PhD,
Cumulative incidence of local recurrence in the whole trial was eight of 357 patients who received short-course Prof B Glimelius PhD);
Department of
radiotherapy, ten of 355 who received short-course radiotherapy with delay, and seven of 128 who received long-
Oncology-Pathology,
course radiotherapy (HR vs short-course radiotherapy: short-course radiotherapy with delay 1∙44 [95% CI 0∙41–5∙11]; Karolinska Institutet,
long-course radiotherapy with delay 2∙24 [0∙71–7∙10]; p=0∙48; both deemed non-inferior). Acute radiation-induced Stockholm, Sweden
toxicity was recorded in one patient (<1%) of 357 after short-course radiotherapy, 23 (7%) of 355 after short-course (H Johansson MSc); Department
of Clinical Sciences, Danderyd
radiotherapy with delay, and six (5%) of 128 patients after long-course radiotherapy with delay. Frequency of Hospital and Ersta Hospital,
postoperative complications was similar between all arms when the three-arm randomisation was analysed (65 [50%] Karolinska Institutet,
of 129 patients in the short-course radiotherapy group; 48 [38%] of 128 patients in the short-course radiotherapy with Stockholm, Sweden
delay group; 50 [39%] of 128 patients in the long-course radiotherapy with delay group; odds ratio [OR] vs short- (M Machado PhD , F Hjern PhD);
Department of Clinical and
course radiotherapy: short-course radiotherapy with delay 0∙59 [95% CI 0∙36–0∙97], long-course radiotherapy with Experimental Medicine,
delay 0∙63 [0∙38–1∙04], p=0∙075). However, in a pooled analysis of the two short-course radiotherapy regimens, the Linköping University,
risk of postoperative complications was significantly lower after short-course radiotherapy with delay than after Linköping, Sweden
short-course radiotherapy (144 [53%] of 355 vs 188 [41%] of 357; OR 0∙61 [95% CI 0∙45–0∙83] p=0∙001). (Prof O Hallböök PhD); and
Department of Surgery, Lund
University, Malmö, Sweden
Interpretation Delaying surgery after short-course radiotherapy gives similar oncological results compared with short- (I Syk PhD)
course radiotherapy with immediate surgery. Long-course radiotherapy with delay is similar to both short-course Correspondence to:
radiotherapy regimens, but prolongs the treatment time substantially. Although radiation-induced toxicity was seen Dr Johan Erlandsson, Centre of
after short-course radiotherapy with delay, postoperative complications were significantly reduced compared with Digestive Diseases P9:03,
Karolinska University Hospital,
short-course radiotherapy. Based on these findings, we suggest that short-course radiotherapy with delay to surgery is SE 17176 Stockholm, Sweden
a useful alternative to conventional short-course radiotherapy with immediate surgery. johan.erlandsson@ki.se

Funding Swedish Research Council, Swedish Cancer Society, Stockholm Cancer Society, and the Regional Agreement
on Medical Training and Clinical Research in Stockholm.

www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4 1

 Articles
Research in context
Evidence before the study
At initiation of the trial, we searched PubMed, MEDLINE, and the
Cochrane Library for all publications in the English language with
a publication date up to April 30, 1998, with search terms,
including rectal cancer, radiotherapy, and timing to surgery. An
updated search was done with same search terms but with a
publication date up to Aug 31, 2016. Several randomised studies
and a meta-analysis had shown that preoperative radiotherapy
reduces the risk of local recurrence by about 50–70% and is more
effective and better tolerated than when given postoperatively. In
many countries, a short radiotherapy course (5 × 5 Gy in 1 week)
with surgery within a week is common practice, as is long-course
radiotherapy (25–28 fractions × 1·8–2 Gy) and surgery after
4–8 weeks, most often without concomitant chemotherapy. A
third option, 5 × 5 Gy with surgery delayed for 4–8 weeks
(short-course radiotherapy with delay) was an alternative, with
potentially fewer postoperative complications than short-course
radiotherapy (with immediate surgery) and tumour regression,
similar to long-course radiotherapy alone or with concomitant
chemotherapy. During the Stockholm III trial period, other studies
confirmed the positive effect of radiotherapy on local recurrence
rates, that adding chemotherapy concomitantly to radiotherapy
further reduced local recurrence rates, and that short-course
radiotherapy (with immediate surgery) is a safe alternative to
chemoradiotherapy and is better tolerated. Further, retrospective
studies have shown that short-course radiotherapy with delay in
patients not fit for chemotherapy concomitantly with
radiotherapy was well tolerated and could result in substantial
tumour regression. However, no randomised trial had compared
short-course radiotherapy versus short-course radiotherapy with
delay or long-course radiotherapy with delay.
Added value of this study
The Stockholm III trial was launched in 1998 to determine the
optimal fractionation of preoperative radiotherapy and timing
Introduction
Preoperative radiotherapy reduces the risk of local
recurrence after surgery for rectal cancer by more than
50%, even with optimised total mesorectal excision
surgery.1–6 Conventionally fractionated long-course
radiotherapy (ie, five fractions of 1·8–2 Gy per week
during 5–6 weeks), most often in combination with
chemotherapy, has been the predominant treatment
in most countries. Short-course radiotherapy (ie, five
fractions of 5 Gy in 1 week [5 × 5 Gy]), and surgery within
the following week has been commonly used in Sweden
and in some other countries in northern and western
Europe. However, the optimal fractionation and timing of
surgery in relation to radiotherapy is still controversial.7
Short-course radiotherapy with surgery delayed for
4–8 weeks (short-course radiotherapy with delay) is an
alternative treatment option that might lead to fewer post-
operative complications and enhance tumour regression,
2 www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4
to surgery for patients with resectable rectal cancer using three
treatment regimens: short-course radiotherapy, short-course
radiotherapy with delay, and long-course radiotherapy with
delay. After a follow-up of a minimum of 2 years, all treatments
were well tolerated and no differences in local or distant
recurrences or overall survival among the three different
radiotherapy regimens were noted. However, the risk of surgical
complications was significantly reduced by delaying surgery.
In both groups where surgery was delayed, radiation-induced
toxicity of grade 3–4 was seen in about 6% of the patients.
Long-course radiotherapy with delay was similar to
short-course radiotherapy with delay, but prolonged treatment
time substantially.
Implications of all the available evidence
During the past three decades the prognosis for patients with
rectal cancer has improved and the incidence of local recurrence
has decreased. The aim of rectal cancer care today must be to
retain low rates of local disease recurrence, reduce the risk of
systemic recurrence, and reduce both acute and long-term
side-effects. The results of this study suggest that short-course
radiotherapy with delay is non-inferior to short-course
radiotherapy, that it is oncologically safe to delay surgery
4–8 weeks after short-course radiotherapy, and that it results in
fewer postoperative complications. A potential disadvantage is
that any preoperative delay will also delay the start of adjuvant
chemotherapy; however, the effect on survival of adjuvant
chemotherapy in patients treated with preoperative
chemotherapy concomitantly with radiotherapy is
controversial. Additionally, chemotherapy for rectal cancer
could be more effective if given before, rather than after,
surgery. Thus, short-course radiotherapy with delay presents
the opportunity to give neoadjuvant chemotherapy during the
interval between radiotherapy and surgery.
which could facilitate surgery. These three regimens,
without concomitant chemotherapy, have never been
compared in a prospective randomised trial and the effects
on tumour response, local recurrence, radiation toxicity,
and postoperative complications are still contentious.
The Stockholm III trial aimed to compare these
three different schedules of radiotherapy (short-course
radiotherapy, short-course radiotherapy with delay, and
long-course radiotherapy with delay) in patients with
primary adenocarcinoma of the rectum. Two preplanned
interim analyses have shown that patients in the short-
course radiotherapy with delay group had a lower
pathological tumour stage, a higher proportion of
patients achieving a complete pathological response, and
a greater degree of tumour regression than did patients
in the short-course radiotherapy group.8,9 Here, we
present the results of the primary outcome, time to local
recurrence, after a minimum follow-up of 2 years.

Methods
Study design and participants
The details of the Stockholm III trial, a multicentre,
randomised, non-blinded, phase 3, non-inferiority trial
have been described previously.8 In summary, patients
scheduled for an open abdominal procedure with a
biopsy-proven primary adenocarcinoma of the rectum,
defined as an adenocarcinoma within 15 cm of the anal
verge, without signs of non-resectability or distant
metastases, and without previous radiotherapy to the
abdominal or pelvic regions, signs of severe ischaemic
disease, or symptoms of severe arteriosclerosis, with no
age restriction, were eligible. All patients gave informed
consent. The trial was approved by the Regional Ethics
Committee at Karolinska Institutet, Stockholm, Sweden
(ethical permission number 98–240, amendment
May 21, 1999) and at all participating hospitals. The
protocol is available online.
Randomisation and masking
Patients were randomly assigned (1:1:1) to short-course
radiotherapy (5 × 5 Gy) with surgery within 1 week
(short-course radiotherapy), short-course radiotherapy
(5 × 5 Gy) with surgery after 4–8 weeks (short-course
radiotherapy with delay), or long-course radiotherapy
(25 × 2 Gy) with surgery after 4–8 weeks (long-course
radiotherapy with delay).
After a protocol amendment on May 21, 1999,
participating hospitals could choose to randomise
patients to just the short-course radiotherapy arms (1:1) or
to all three arms. Randomisation was done by telephone
by the Regional Cancer Centre in Stockholm, Sweden,
after assessing inclusion and exclusion criteria. Computer
generated randomisation lists were constructed by use of
permuted block technique (block size of six for the three-
arm randomisation, block size of four for the two-arm
randomisation) and stratified by participating centre.
Investigators and patients were not masked to treatment.
Procedures
Preoperative staging included assessment of local tumour
extent and distant metastases. During the first years,
1998–2003, staging method varied between hospitals, but
after 2003, pelvic MRI and CT of the liver and chest
became standard at all hospitals, replacing the previous
use of chest radiography and ultrasound of the liver.
Radiotherapy was given with a three-beam or
four-beam box technique, including the primary tumour
and primary and secondary lymph nodes in the pelvis,
as described previously.8 Patients randomly assigned to
one of the two short-course schedules received
radiotherapy with five fractions of 5 Gy to a total dose of
25 Gy during 5 consecutive days, preferably Monday to
Friday. Patients in the long-course radiotherapy with
delay group received a total dose of 50 Gy administered
as 25 fractions of 2 Gy. No concomitant chemotherapy
was given.
www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4
Articles
Tumours were resected by anterior resection (ie,
sphincter-preservation surgery), abdominoperineal
excision, or Hartmann’s procedure using the total
mesorectal excision technique. Patients randomly
assigned to the short-course radiotherapy group were to
have surgery 1–7 days after radiotherapy. In the short-
course radiotherapy delay group and the long-course
radiotherapy delay group, the surgery was to be done
28–56 days after radiotherapy. Bowel preparation,
prophylactic antibiotics, and prophylactic anti-
thrombotic drugs were given according to local routines
at participating centres.
Follow-up, which included local recurrence, distant
metastases, and adverse events, using chest radiograph or
CT scan of the chest and CT scan of abdomen—MRI was
used if there was a suspicion of local recurrence and
endoscopy was used at the discretion of the treating
physician—was recommended, according to protocol, at For the protocol see http://ki.se/
3 months, 6 months, and 12 months after surgery, and en/mmk/stockholm-iii-trial
yearly thereafter, although follow-up according to the
national guidelines with a minimum follow-up at 1 year
and 3 years was allowed. All follow-up was done in person
or by telephone contact with the patient and recorded in
the registry.
Data for all patients with rectal cancer are reported
continuously to the Swedish ColoRectal Cancer
Register (SCRCR), a nationwide, validated registry.10
The registry contains information about patient and
tumour characteristics, neoadjuvant therapy, short-term
and long-term complications, recurrences, and death.
Data for adjuvant chemotherapy were also available in
the registry starting from Jan 1, 2007. A diagnosis of
local recurrence was confirmed and reported by the
patient-responsible physician to the SCRCR. Local
recurrence was defined as tumour growth below the
level of the sacral promontory, related to the previous
rectal cancer, and diagnosed radiographically with MRI,
CT, or both, or clinically (preferably with histological
confirmation). Distant metastasis was defined as a
radiological or clinical finding (CT, or MRI, or both)
consistent with rectal cancer metastases outside the
irradiation field, preferably with histological
confirmation. Data from the registry regarding acute
radiation toxicity, postoperative complications, and
recurrences were validated with the patients’ medical
charts. Sphincter preservation rate, reoperations, and
local and distant recurrences were recorded in the
registry. Data for late toxicity and complications were
available from the SCRCR and were obtained without
validation with the patients’ medical charts. All patients
are reported at 1 year, 3 years, and 5 years after surgery
according to the protocol. If an earlier event occurred
they are reported in between. If patients did not register
at 3 years, we asked the patient’s responsible physician
to report any events. The last day of follow-up was set to
March 30, 2015 (censoring date), when all patients had
been followed for at least 2 years.
3
 Articles
Outcomes
The primary endpoint was time to local recurrence
calculated from the date of randomisation to the date of
local recurrence. Secondary endpoints were overall
survival calculated from the date of randomisation until
death from any cause or emigration, frequency of
postoperative mortality (defined as any mortality within
30 days of surgery or within hospital stay after primary
surgery), frequency of postoperative complications,
frequency of reoperations, frequency of late
complications, radiation toxicity, and frequency of
sphincter-preserving surgeries (anterior resections). In a
protocol amendment (May 21, 1999) quality of life
was added as a secondary endpoint; these results will
be reported elsewhere. Distant metastases-free survival
See Online for appendix (defined as time from randomisation to distant
metastases) and recurrence-free survival (defined as time
from randomisation to local or distant recurrence) were
considered exploratory endpoints for post-hoc analyses.
The original protocol listed frequency of defunctioning
stoma as a secondary endpoint, but after the completion
of the study protocol the Swedish national guidelines
stated that it is mandatory to deviate any rectal
anastomosis with a stoma after radiotherapy, thus this
endpoint could not be assessed. Intercurrent death was
defined as death without signs of recurrence of rectal
cancer. Postoperative complications were defined as any
surgical, cardiovascular, infectious, or neurological
complication noted within 30 days after surgery. Surgical
complications were defined as any surgical-site infection,
deep infection, anastomotic leak, postoperative bleeding,
stoma-related complication, wound dehiscence, or other
surgical related adverse outcome. Classification of post-
operative complications according to Clavien-Dindo was
not possible during the first years (1998–2007) of the trial;
instead total number of events and frequency of
reoperations are presented. Reoperation was defined as a
surgical intervention requiring anaesthesia due to a
postoperative complication. Late complications and late
radiotherapy toxicity were identified from a minimum of
30 days after surgery and defined as bowel obstruction,
pelvic abscess, anal incontinence, anastomotic or stomal
complication, incisional hernia, sexual dysfunction,
urinary problem, pelvic insufficiency fracture, or other
(fistula [most common rectovaginal fistula], long-lasting
pelvic infection, and severe perineal or pelvic pain), and
reported by the responsible surgeon.
Analysis of acute radiation toxicity was not stated as a
secondary outcome in the original protocol; however,
this was considered an important factor in relation to
postoperative complications and was analysed post hoc
as a matter of safety. Acute radiation toxicity was
defined as unplanned patient admission to hospital
with symptoms of radiation side-effects, corresponding
to the Radiotherapy Oncology Group (RTOG) scales 3–4,
between radiotherapy initiation and the date of
surgery.11
4 www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4
Statistical analysis
The study was designed as a non-inferiority trial. Treatment
groups were deemed non-inferior if the upper limit of a
two-sided 90% CI for the hazard ratio (HR) did not
exceed 1∙7, corresponding to a one-sided test at the
significance level of 0∙05. The primary comparison was of
short-course radiotherapy with either short-course
radiotherapy with delay or long-course radiotherapy with
delay. With a cumulative incidence of local recurrence
at 15%, an estimation based on studies published before
trial initiation, a sample size of 840 patients was determined
to provide a power of 80%. Because it became apparent
soon after study commencement that local recurrence
would be much lower than anticipated, a revised power
calculation was done (protocol amendment, May 21, 1999)
and additional secondary endpoints were defined (appendix
p 1). In summary, the primary endpoint was not changed,
and the trial was re-powered to deem non-inferiority at an
HR not exceeding 2∙5 or 3∙6, depending on the frequency
of local recurrence (appendix p 1). A pre-planned interim
analysis done in 303 patients enrolled between Oct 1, 1998,
and Dec 31, 2005, did not identify any differences in
radiation toxicity or postoperative complications between
all treatment groups and patient enrolment continued.8
Cumulative incidences of local recurrence and distant
metastases were estimated with non-parametric methods
taking competing risks into account. Cumulative
incidences of local recurrence or distant metastases were
reported as first event and as any event. Local recurrence,
distant metastases, and death were included as
competing risks to each other. Death was not deemed to
have any competing risk. Effect of treatment on endpoints
was estimated with proportional hazards regression,
stratified by participating centre. Effect of treatment on
recurrence- free survival and overall survival were
analysed with Kaplan-Meier curves. Results are presented
as HRs with 95% CIs. For the primary outcome, time to
local recurrence results were presented with a 90% CI,
corresponding to the one-sided significance level of
0∙05. p values were calculated with the log-rank test.
Odds ratios (OR) for prespecified endpoints, post-
operative complications, reoperations, and radiation
toxicity, with p values were calculated by logistic
regression. Dichotomous variables, including post-
operative complications and late toxicity, were analysed
with χ² test or Fisher’s exact test when appropriate.
Patients in the three-arm randomisation were analysed
separately from those in the two-arm randomisation for
all endpoints. Additionally, patients randomly assigned
to short-course radiotherapy and short-course radio-
therapy with delay in both randomisation groups were
pooled and analysed together for all endpoints in a
prespecified analysis. To ensure that the pooling of
patients to a short radiotherapy comparison was
adequate, an interaction analysis of randomisation
design and radiotherapy treatment was done on the
oncological outcomes (appendix p 6).
 Articles

Because there was an imbalance regarding tumour
height from anal verge, and subsequently surgical
technique between the treatment groups, and to assess the
randomisation process, a multivariate model including
age, sex, tumour height, and surgical technique was also
analysed for recurrence-free survival as a post-hoc analysis.
We did a post-hoc analysis of all oncological outcomes
using date of surgery as the starting point instead of
randomisation date to account for differences in follow-up
time. Frequency of adjuvant chemotherapy was compared
between treatment groups in a post-hoc analysis. All
analyses were done according to intention to treat. For the
safety assessment of radiation toxicity we also did two post-
hoc analyses, one per-protocol analysis (ie, only patients
without protocol deviations were included) and one
as-treated assessment (ie, patients were analysed
according to the actual radiotherapy course they received).
Data were analysed with Stata version 14.
The trial was registered with Clinicaltrials.gov, number
NCT00904813.
Role of the funding source
The funders had no role in the study design, data
collection, data analysis, data interpretation or writing of
the report. All authors had full access to all the data in the
trial. JE and AM had final responsibility for the decision
to submit for publication.
Results
Between Oct 5, 1998, and Jan 31, 2013, 840 eligible
patients were recruited from 18 hospitals in Sweden
(appendix p 9). 385 patients were randomly assigned
between short-course radiotherapy, short-course radio-
therapy with delay, and long-course radiotherapy with
delay in the three-arm randomisation and 455 patients
between short-course radiotherapy and short-course
radiotherapy with delay in the two-arm randomisation
(figure 1). Hospitals that chose to participate in only the
two-arm randomisation generally did so for logistical
reasons, such as insufficient radiotherapy capacity during
parts of the year, patients not consenting to be randomised
to long-course radiotherapy with delay, or hospital
preference to take part only in the two-arm randomisation.
In the three-arm randomisation, 129 patients were
randomly assigned to short-course radiotherapy, 128 to
short-course radiotherapy with delay, and 128 to long-
course radiotherapy with delay. In the two-arm
randomisation, 228 patients were randomly assigned to
short-course radiotherapy and 227 were assigned to short-
course radiotherapy with delay. 74 (9%) patients had a

8122 eligible patients

7282 excluded
4870 did not meet inclusion criteria
2171 not asked to participate
216 administrative reasons
25 other

385 assigned to three-arm 455 assigned to two-arm

randomisation randomisation

129 allocated to SRT 128 allocated to SRT-delay 128 allocated to LRT-delay 228 allocated to SRT 227 allocated to SRT-delay

127 5 × 5 Gy radiotherapy 128 5 × 5 Gy radiotherapy 16 5 × 5 Gy radiotherapy 225 5 × 5 Gy radiotherapy 226 5 × 5 Gy radiotherapy

2 no radiotherapy
123 <1 week to surgery
1 2–3 weeks to surgery
3 4–8 weeks to surgery
2 <1 week to surgery
2 2–3 weeks to surgery
121 4–8 weeks to surgery
3 >9 weeks to surgery
1 no surgery*
110 25 × 2 Gy radiotherapy
2 no radiotherapy
7 <1 week to surgery
5 2–3 weeks to surgery
102 4–8 weeks to surgery
10 >9 weeks to surgery
3 no surgery†
1 other radiotherapy‡
2 no radiotherapy
210 <1 week to surgery
5 2–3 weeks to surgery
7 4–8 weeks to surgery
4 >9 weeks to surgery
1 no radiotherapy
8 <1 week to surgery
0 2–3 weeks to surgery
214 4–8 weeks to surgery
4 >9 weeks to surgery

129 analysed for all 128 analysed for all 128 analysed for all 228 analysed for all 227 analysed for all
endpoints endpoints endpoints endpoints endpoints

Figure 1: Trial profile

For the protocol violations, some patients were counted twice as it is events that is reported (see appendix p 4). SRT=short-course radiotherapy (5 × 5 Gy with surgery
within 1 week). SRT-delay=short-course radiotherapy (5 × 5 Gy with surgery after 4–8 weeks). LRT-delay=long-course radiotherapy (25 × 2 Gy with surgery after
4–8 weeks). Interval to surgery is time from end of radiotherapy until surgery. *Patient with local excision who had an interval to surgery. †Two patients without surgery
and one patient with a local excision. ‡Patient received 4 Gy x 5 Gy.

www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4 5

 Articles

short-course radiotherapy group, 28 (13 %) of 216 patients

Three-arm randomisation Two-arm randomisation
in the short-course radiotherapy group with delay, and 10
SRT SRT-delay LRT-delay SRT SRT-delay (19%) of 53 patients in long-course radiotherapy with delay
(n=129) (n=128) (n=128) (n=228) (n=227) group. 63 (34%) of 188 patients had (y)pStage (pathology
Age (years) 67 (62–74) 67 (62–75) 66 (61–73) 67 (61–74) 67 (61–74) stage after neoadjuvant treatment) III (two patients in (y)
Sex pStage I and seven patients had (y) pStage III); in a post-
Male 81 (63%) 79 (62%) 73 (57%) 137 (60%) 134 (59%) hoc analysis, frequency of adjuvant chemotherapy was not
Female 48 (37%) 49 (38%) 55 (43%) 91 (40%) 93 (41%)
significantly different between the groups. Median follow-
up time for all patients was 5∙2 years (IQR 3∙7–6∙1; range
Height from anal verge
2∙0–14∙6).
0–5 cm 50 (39%) 57 (45%) 31 (25%) 78 (34%) 68 (30%)
In the three-arm randomisation, the median interval
6–10 cm 49 (38%) 49 (39%) 60 (48%) 91 (40%) 95 (42%) between start of radiotherapy and surgery was 8 days
11–15 cm 30 (23%) 21 (17%) 35 (28%) 58 (26%) 63 (28%) (IQR 7–10) in the short-course radiotherapy group, 47 days
Type of surgery (41–55) in the short-course radiotherapy with delay group,
Anterior resection 79 (61%) 68 (53%) 93 (72 %) 139 (61%) 136 (60%) and 76 days (70–84) in the long-course radiotherapy with
Abdominal perineal excision 47 (36%) 53 (41%) 24 (19%) 75 (33%) 79 (35%) delay group. No significant difference between treatment
Hartmann’s 3 (2%) 6 (5%) 8 (6%) 14 (6%) 12 (5%) groups was noted for cumulative incidence of local
Local excision 0 (0%) 1 (1%) 0 (0%) 0 (0%) 0 (0%)
recurrence, cumulative incidence of distant metastases,
overall survival, or cumulative incidence of intercurrent
No resection 0 (0%) 0 (0%) 3 (2%) 0 (0%) 0 0(%)
death (table 2, figure 2). In patients with any local
ypStage
recurrence the median time from date of randomisation to
I* 38 (29%) 55 (43%) 37 (29%) 58 (26%) 83 (36%) local recurrence was 28∙3 months (min 20∙7–max 62∙2) in
II 43 (33%) 31 (24 %) 46 (37%) 75 (33%) 55 (24%) short-course radiotherapy group, 22∙1 months (15∙5–34∙3)
III 48 (37%) 31 (24%) 37 (30%) 86 (38%) 76 (34%) in short-course radiotherapy with delay group, and 33∙3
IV 0 (0%) 7 (6%) 5 (4%) 7 (3%) 6 (3 %) months (17∙8–114∙3) in long-course radiotherapy with
Stage x 0 (0%) 3 (2%) 1 (1%) 1 (1%) 6 (3%) delay group (appendix p 5). χ² statistics and degrees of
freedom are shown in the appendix (p 6). 5-year overall
Data are median (IQR) or n (%). SRT=short-course radiotherapy 5 × 5 Gy and surgery within 1 week. SRT-delay=short-
survival was 73% (95% CI 64–80) for short-course
course radiotherapy, 5 × 5 Gy and surgery after 4–8 weeks. LRT-delay=long-course radiotherapy, 25 × 2 Gy and surgery
after 4–8 weeks. ypStage=pathological stage after neoadjuvant treatment. Stage x=stage unknown. *Includes patients radiotherapy, 76% (67–83) for short-course radiotherapy
with complete response. with delay, and 78% (70–84) for long-course radiotherapy
with delay (figure 2). In a post-hoc analysis, 5-year
Table 1: Baseline characteristics, type of surgery, and postoperative stage of the 840 patients by
allocated treatment recurrence-free survival was 65% (95% CI 56–73) after 44
recurrence events in the short-course radiotherapy group,
64% (54–71) after 45 recurrence events in the short-course
protocol violation. Deviations from protocol were no radiotherapy with delay group, and 65% (56–73) after 44
radiotherapy (seven patients [<1%]), different fractionation recurrence events in the long-course radiotherapy with
schedule than random assignment (17 patients [2%]), delay group (figure 2). Table 2 shows postoperative and
incorrect interval between radiotherapy and surgery surgical complications and reoperations at the end of
(61 patients [7%]), no abdominal surgery (three patients follow-up. In a post-hoc analysis, after adjusting for
<1%), or a combination of the above (three patients [<1% covariates in a multivariate model, recurrence-free survival
all in long-course radiotherapy with delay); two patients remained similar between treatment groups (for short-
with deviant radiotherapy (other than randomised to) and course radiotherapy (ref) vs short-course radiotherapy with
no surgery, and one patient with no radiotherapy and no delay HR 0∙98 [95% 0∙67–1∙44]; vs long-course
surgery; figure 1). Baseline patient characteristics were radiotherapy with delay 1∙01 [0∙69–1∙46]).
well balanced, except in the three-arm randomisation Patients randomly assigned to short-course radiotherapy
where there was a higher number of patients with a greater with immediate or delayed surgery were pooled from the
tumour height in the long-course radiotherapy with delay three-arm and two-arm randomisations, resulting in
group compared with other groups, which could account 357 patients in the short-course radiotherapy group and
for the higher number of patients who had an anterior 355 patients in the short-course radiotherapy with delay
resection in that group (table 1). No patient had distant group. Characteristics of these patients are shown in the
metastases detected at the time of randomisation; however, appendix (p 5). Median time between the start of
15 patients (2%) had metastases detected at the time of radiotherapy and surgery was 8 days (IQR 7–10) for short-
surgery and remained in the analyses. 489 (58%) of course radiotherapy and 45 days (42–51) for short-course
840 patients were included after Jan 1, 2007, when adjuvant radiotherapy with delay.
chemotherapy data became available in the SCRCR The cumulative incidences of local recurrence, distant
registry (appendix p 2). Adjuvant chemotherapy was given metastases, and intercurrent death did not differ
to 72 (15%) of 489 patients; 29 (13%) of 220 patients in the significantly between the two treatment groups (table 3,

6 www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4

 Articles

figure 3). In patients with any local recurrence the

Three-arm randomisation p value
median time from date of randomisation to local
recurrence was 33∙4 months (min 18∙2–max 62∙2) in SRT (n=129) SRT-delay (n=128) LRT-delay (n=128)
the short-course radiotherapy group and 19∙3 months Local recurrence
(8∙5–39∙5) in the short-course radiotherapy with delay As the first recurrence event 3 1 4 ..
group (appendix p 5). χ² statistics and degrees of freedom HR (90% CI) 1·00 (ref) 0·38 (0·06–2·56) 1·22 (0·33–3·45) 0·52*
from the incidence and survival curves are shown in the As any event 3 4 7 ..
appendix (p 5). Recurrence-free survival and overall HR (90% CI) 1·00 (ref) 1·44 (0·41–5·11) 2·24 (0·71–7·10) 0·48*
survival also did not differ significantly between Distant metastases
treatment groups (table 3). At 5 years, overall survival was As the first recurrence event 29 38 35 ..
76% (95% CI 70–80) for short-course radiotherapy and HR (95% CI) 1·00 (ref) 1·45 (0·89–2·37) 1·25 (0·76–2·04) 0·33*
77% (72–82) for short-course radiotherapy with delay; As any event 30 38 35 ..
recurrence-free survival for short-course radiotherapy HR (95% CI) 1·00 (ref) 1·40 (0·84–2·18) 1·20 (0·74–1·96) 0·40*
after 115 events was 65% (95% CI 60–70) and for short- Survival
course radiotherapy with delay after 106 events was Deaths 51 43 49 ..
68% (62–73; figure 3). Recurrence-free survival remained Number of patients with any lethal 61 54 58 ..
similar in both groups after post-hoc adjustment in the event
multivariate model (short-course radiotherapy [ref ] vs Overall survival HR (95% CI) 1·0 (ref) 0·81 (0·53–1·24) 0·94 (0·63–1·40) 0·62*
short-course radiotherapy with delay HR 0∙92 [95% CI Recurrence-free survival HR (95% CI) 1·0 (ref) 0·93 (0·64–1·35) 0·99 (0·68–1·42) 0·92*
0∙72–1∙18]). Number of intercurrent deaths 29 15 19 ..
The upper boundary of HR 1∙7 could not be excluded Intercurrent death HR (95% CI) 1·00 (ref) 0·46 (0·24–0·90) 0·70 (0·38–1·26) 0·06*
when analysing the cumulative incidence of local Complications
recurrence in both the three-arm randomisation and the Any postoperative complication 65 (50%) 48 (38%) 50 (39%) ..
pooled short radiotherapy course comparison; however, OR (95% CI) 1·0 (ref) 0·59 (0·36–0·97) 0·63 (0·38–1·04) 0·075†
the results are in line with the revised power calculation Any surgical complication 40 (31 %) 33 (26%) 30 (23%) ..
and an HR of 3∙6 could be excluded, which indicates that OR (95% CI) 1·0 (ref) 0·77 (0·49–1·33) 0·68 (0·39–1·18) 0·38†
short-course radiotherapy with delay is non-inferior to Reoperation 11 (11%) 7 (7%) 7 (7 %) ..
short-course radiotherapy (appendix p 1). Short-course OR (95% CI) 1·0 (ref) 0·64 (0·24–1·73) 0·63 (0·23–1·68) 0·56†
radiotherapy with delay was also deemed non-inferior to
short-course radiotherapy for all remaining oncological Data are n (%) if not otherwise specified. SRT=short-course radiotherapy 5 × 5 Gy, surgery within 1 week.
SRT-delay=short-course radiotherapy 5 × 5 Gy, with a delay of 4–8 weeks to surgery. LRT-delay=long-course radiotherapy
outcomes (appendix p 3). 25 × 2 Gy, with a delay of 4–8 weeks to surgery. HR=hazard ratio. OR=odds ratio. Postoperative complication was defined as
When patients randomly assigned to short-course any cardiovascular, infectious, neurological, or surgical complication noted within 30 days after surgery. Surgical
radiotherapy or short-course radiotherapy with delay complication was defined as any surgical site infection, deep infection, anastomotic leak, postoperative bleeding, stoma
related complications, wound dehiscence, or other surgical complication. *Overall p value, non-parametric Cox test
were analysed separately according to whether they were stratified by hospital. †Logistic regression with X2 test.
part of the three-arm or two-arm randomisations, no
differences in local recurrence, distant metastases, Table 2: Oncological outcomes and postoperative complications by allocated treatment in the three-arm
randomisation at the end of follow-up
overall survival, or recurrence-free survival were noted
(appendix p 6). However, in the three-arm randomisation,
patients randomly assigned to short-course radiotherapy (seven [23%] of 30 patients). No patient died due to
had a higher probability of intercurrent death compared radiation toxicity during the interval from start of
with those assigned to short-course radiotherapy, radiotherapy to surgery.
although this was not noted in the two-arm randomisation In the three-arm randomisation, no patient in the short-
(appendix p 6). course radiotherapy group was admitted to hospital due
30-day mortality for all patients was less than 1% to radiation toxicity, but seven (6%) of 128 patients in the
(six patients); deaths occurred in two patients after short- short-course radiotherapy with delay group and six (5%)
course radiotherapy, three after short-course radiotherapy of 128 patients in the long-course radiotherapy with delay
with delay, and one after long-course radiotherapy with group were admitted; this difference was not significant
delay. Causes of death were myocardial infarction in (OR 0∙84 [95% CI 0·28–2·58]; c²=0∙77). In the pooled
four (<1%; two in short-course radiotherapy with delay short-course radiotherapy comparison, one (<1%) of
and two in short-course radiotherapy), respiratory failure 357 patients was admitted due to radiation toxicity in
in one (<1%; one in short-course radiotherapy with the short-course radiotherapy group and 23 (7%) of
delay), and unknown in one (<1%; in long-course 355 patients were admitted for radiation toxicity in the
radiotherapy). None of the causes of death were judged to short-course radiotherapy with delay group (OR 24∙67
be a consequence of allocated treatment. 30 patients (4%) (95% CI 3∙31–183∙72]; p<0∙0001). Post-hoc analyses
were admitted to hospital due to acute radiation toxicity. comparisons of acute radiation toxicity in the per-protocol
The two most common reasons for admission were and as-treated populations showed similar results as the
diarrhoea (11 [37%] of 30 patients) and abdominal pain intention-to-treat analyses (appendix p 6).

www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4 7

 Articles

40 SRT 40
SRT-delay
35 LRT-delay 35
30 30
25 25
20 20
15 15
A 10 B 10
100 p=0·48 100 p=0·40

Cumulative incidence of
Cumulative incidence of

distant metastases (%)

5 5
local recurrence (%)

80 80
0 0
60 0 2 4 6 8 10 60 0 2 4 6 8 10
40 40
20 20
0 0
Number at risk 0 2 4 6 8 10 0 2 4 6 8 10
(censored)
SRT 129 (0) 101 (28) 82 (46) 36 (91) 18 (109) 12 (115) 129 (0) 101 (11) 82 (22) 36 (65) 18 (82) 12 (88)
SRT-delay 128 (0) 101 (25) 77 (48) 36 (89) 20 (105) 17 (108) 128 (0) 101 (5) 77 (19) 36 (55) 20 (71) 17 (74)
LRT-delay 128 (0) 98 (28) 75 (48) 34 (89) 20 (103) 9 (113) 128 (0) 97 (9) 74 (25) 34 (62) 20 (74) 9 (85)

C D
Recurrence-free survival (%)

100 100
Overall survival (%)

80 80
60 60
40 40
20 20
p=0·92 p=0·61
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Number at risk Time since randomisation (years) Time since randomisation (years)
(censored)
SRT 129 (0) 101 (0) 82 (8) 36 (44) 18 (56) 12 (60) 129 (0) 112 (0) 92 (11) 42 (51) 19 (65) 12 (70)
SRT-delay 128 (0) 101 (0) 77 (11) 36 (45) 20 (59) 17 (62) 128 (0) 114 (0) 92 (14) 43 (52) 22 (68) 18 (72)
LRT-delay 128 (0) 97 (1) 74 (13) 34 (47) 20 (55) 9 (65) 128 (0) 110 (2) 89 (15) 41 (53) 23 (62) 11 (74)

Figure 2: Local recurrence (A), distant metastases (B), recurrence-free survival (C), and overall survival (D) in the three-arm randomisation with a minimum of 2 years of follow-up
SRT=short-course radiotherapy (5 × 5 Gy with surgery within 1 week). SRT-delay=short-course radiotherapy (5 × 5 Gy with surgery after 4–8 weeks). LRT-delay=long-course radiotherapy (25 × 2 Gy with
surgery after 4–8 weeks).

The most common postoperative complication was (55 [7%]; appendix p 7). No significant differences between
surgical-site infection (154 [18%] of 840 patients; 77 (22%) groups were noted in the three-arm randomisation or in
of 357 patients in the short-course radiotherapy group, the short-course radiotherapy comparison (appendix p 7).
59 (17%) of 355 patients in the short-course radiotherapy The post-hoc analyses of local recurrence, distant
with delay group, 18 (14%) of 128 patients in metastases, overall survival, and recurrence-free survival
the long-course radiotherapy with delay group. This did using the date of surgery instead of the date of random-
not differ between the groups in either the three-arm isation as the time of origin did not change the results in
randomisation or in the pooled short-course radiotherapy the three-arm randomisation analysis set or the short-
comparison. Postoperative and surgical complications course radiotherapy combined population (appendix p 8).
did not differ between treatment groups in the three-
arm randomisation (table 2) but significantly fewer Discussion
surgical and overall complications were reported after In this multicentre, randomised, non-inferiority trial, no
short-course radiotherapy with delay than after short- significant differences between three different preoperative
course radiotherapy in the short-course radiotherapy radiotherapy regimens for rectal cancer were observed
comparison (table 3). In 87 patients (10%) a complication regarding time to local or distant recurrence, recurrence-
required a second operation within the same hospital free survival, or overall survival. By delaying surgery for
stay; the frequency of reoperation was not significantly 4–8 weeks after the end of short-course radiotherapy, a
different between treatment groups in either the significantly lower frequency of postoperative comp-
three-arm randomisation or the pooled comparison lications was reported; however, radiation toxicity required
(tables 2, 3). admission to hospital in about 7% of these patients.
The most common late complication was bowel The strength of this study is the randomised, controlled
obstruction (90 [11%] of 840 patients) and pelvic abscesses design. The initial study protocol included three arms

8 www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4

 Articles

(short-course radiotherapy, short-course radiotherapy with

SRT (n=357) SRT-delay (n=355) p value
delay, and long-course radiotherapy with delay). Due to
insufficient resources for radiotherapy at certain hospitals, Local recurrence
and concerns about the long-course radiotherapy-delay As the first recurrence event 7 6 ..
schedule, the possibility of a two-armed randomisation HR (90% CI) 1·00 (ref) 0·91 (0·36–2·27) 0·59*
was introduced as a protocol amendment after about a As any event 8 10 ..
year. This design alteration resulted in an imbalance of HR (90% CI) 1·00 (ref) 1·30 (0·59–2·85) 0·58*
patients, with fewer patients enrolled in the long-course Distant metastases
radiotherapy with delay group than in the short-course As the first recurrence event 77 77 ..
radiotherapy and short-course radiotherapy with delay HR (95% CI) 1·00 (ref) 1·02 (0·74–1·40) 0·91*
groups. To address this imbalance, and any potential As any event 80 79 ..
selection bias, analyses were done in the three-arm HR (95% CI) 1·00 (ref) 1·00 (0·73–1·38) 0·98*
randomisation population in accordance with the initial Survival
protocol, as well as in the population of patients included Deaths 110 108 ..
in either of the short-course radiotherapy groups from Overall survival HR (95% CI) 1·00 (ref) 0·90 (0·70–1·15) 0·46*
both randomisations. A well known disadvantage of Number of patients with any event 141 130 ..
randomised trials is that it can take a long time to complete, Recurrence-free survival HR (95% CI) 1·00 (ref) 0·90 (0·69–1·18) 0·39*
which was the case with this study. During the prolonged Number of patients with intercurrent death 57 47 ..
study period, the local recurrence rate on a population Intercurrent death HR (95% CI) 1·00 (ref) 0·73 (0·49–1·10) 0·13*
basis decreased from 15% to about 5%.12,13 This improved Complications
local control was anticipated after about a year of patient Any postoperative complication 188 (53%) 144 (41%) ..
inclusion and a revision of the power calculations and OR (95% CI) 1·00 (ref) 0·61 (0·45–0·83) 0·001†
outcomes was done as a protocol amendment; however, Any surgical complication 128 (36%) 100 (28%) ..
the primary endpoint and sample size was not adjusted. OR (95% CI) 1·00 (ref) 0·70 (0·51–0·96) 0·03†
In the beginning of the study, patients with early-stage Reoperation 43 (15%) 37 (14%) ..
tumours (T1–T2) could be included, provided abdominal OR (95% CI) 1·00 (ref) 0·88 (0·55–1·41) 0·59†
surgery was planned, but since the implementation of
MRI in 2003 there was a shift in treatment indications Data are n (%) if not otherwise specified. SRT=short-course radiotherapy 5 × 5 Gy surgery within 1 week.
SRT-delay=short-course radiotherapy, 5 × 5 Gy, with a delay of 4–8 weeks to surgery. HR=hazard ratio. OR=odds ratio.
for radiotherapy based on the concept “good”, “bad”, and Postoperative complication was defined as any cardiovascular, infectious, neurologic, or surgical complication
“ugly” tumours.14 According to this concept, “good” observed within 30 days after surgery. Surgical complication was defined as any surgical site infection, deep
tumours are treated with surgery alone, “bad” tumours infection, anastomotic leak, postoperative bleeding, stoma related complications, wound dehiscence, or other
surgical complication. *Cox regression stratified by including centre, log-rank test. †Logistic regression with χ² test.
with radiotherapy and surgery, and “ugly” tumours with
chemoradiotherapy and surgery. Thus, during most of Table 3: Oncological outcomes and postoperative complications by allocated treatment in the pooled
the study period, only patients with “bad” or “intermediate short-course radiotherapy comparison at the end of follow-up
risk” tumours were included in the trial. Although
25–30% of the patients had pathological stage I disease,
with a low risk of local relapse, the stage distribution was Our trial shows similar oncological and safety outcomes
not fundamentally different from other radiotherapy or between short-course radiotherapy and short-course
chemoradiotherapy trials in rectal cancer.5,15–17 radiotherapy with delay. Short-course radiotherapy with
In this trial, outcomes from long-course radiotherapy delayed surgery has also previously been reported to be
with delay were similar to those with short-course radio- well tolerated and to induce complete pathological
therapy with delay in the three-arm randomisation. Since responses.9,23–26 Tumour regression in all patients from the
the trial was initiated, several studies have shown that Stockholm III trial are currently being reassessed with
whenever long-course radiotherapy with delay is con- regard to downstaging and regression and data will be
sidered it should be given with concomitant chemotherapy presented separately. Because of the very short interval
to improve local control.18,19 Previous trials comparing between radiotherapy and surgery in the short-course
short-course radiotherapy (with immediate surgery) to radiotherapy group, the radiation toxicity is probably
conventional long-course chemoradiotherapy have not obscured by early postoperative complications, therefore
shown any significant differences regarding local the relative frequency of toxicity should be interpreted with
recurrence and distant metastases incidence, survival, or caution. In the groups with a delay to surgery, about 6% of
treatment-related complications.17,20 In chemoradiotherapy, patients developed estimated grade 3–4 radiation-induced
a delay before surgery of 6–8 weeks after radiotherapy is toxicity. We have made a thorough review of all patient
standard, and a meta-analysis indicated that even a longer medical charts and every unplanned hospital admission
delay is safe.21 However, a recently published randomised that could be linked to radiation toxicity was recorded as an
trial22 from the French GRECCAR group showed that event, which could explain a rather high incidence.
surgical morbidity was higher when surgery was delayed However, this number is still considerably lower than the
for 11 weeks than for 7 weeks after chemoradiotherapy. level of radiation toxicity after chemoradiotherapy in other

www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4 9

 Articles

40 SRT 40
35 SRT-delay 35
30 30
25 25
20 20
A 15 B 15
100 p=0·58 100 p=0·98
10 10

Cumulative incidence of
Cumulative incidence of

distant metastases (%)

local recurrence (%)

80 5 80 5
0 0
60 60
0 2 4 6 8 10 0 2 4 6 8 10
40 40

20 20

0 0
0 2 4 6 8 10 0 2 4 6 8 10
Number at risk
(censored)
SRT 357 (0) 290 (66) 199 (153) 71 (279) 35 (315) 22 (328) 357 (0) 290 (20) 200 (87) 71 (208) 35 (243) 22 (256)
SRT-delay 355 (0) 289 (61) 195 (151) 84 (262) 45 (301) 31 (315) 355 (0) 290 (18) 195 (92) 84 (194) 45 (232) 31 (246)

C D
100 100
Recurrence-free survival (%)

Overall survival (%)

80 80

60 60

40 40

20 20
p=0·39 p=0·46
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Number at risk Time since randomisation (years) Time since randomisation (years)
(censored)
SRT 357 (0) 290 (1) 199 (59) 71 (163) 35 (192) 22 (200) 357 (0) 326 (3) 231 (70) 86 (186) 37 (221) 22 (231)
SRT-delay 355 (0) 289 (0) 195 (66) 84 (160) 45 (193) 31 (203) 355 (0) 320 (0) 220 (76) 96 (177) 50 (213) 32 (224)

Figure 3: Local recurrence (A), distant metastases (B), recurrence-free survival (C), and overall survival (D) in the pooled short-course radiotherapy comparison with a minimum of 2 years of follow-up
SRT=short-course radiotherapy (5 × 5 Gy with surgery within 1 week). SRT-delay=short-course radiotherapy (5 × 5 Gy with surgery after 4–8 weeks). LRT-delay=long-course radiotherapy (25 × 2 Gy with
surgery after 4–8 weeks).

trials.18,20,27 Because local control has improved substantially
with modern neoadjuvant treatments and optimised total
mesorectal excision surgery, reduction of treatment-related
side-effects like postoperative complications and acute and
late radiation toxicity is important. Long-term data from
this study regarding quality of life (after a minimum
follow-up of 4 years) will be published separately.
As shown in this and other studies, distant metastasis is
now the major cause of rectal cancer relapse.4,12,13 Studies
indicate that a delayed start of adjuvant chemotherapy due
to postoperative complications might have a negative
impact on survival.28 A concern with delaying surgery after
radiotherapy is that it will have a similar effect as delayed
adjuvant treatment. In this trial, a small number of
patients were treated with adjuvant chemotherapy, thus
why it is not possible to exclude a negative influence on
survival. However, the benefit of adjuvant chemotherapy
in patients with rectal cancer treated with radiotherapy or
chemoradiotherapy preoperatively is highly contro-
versial.29,30 On the other hand, a possible benefit of short-
course radiotherapy with delay is that upfront
chemotherapy can be given to patients with a high risk of
distant metastases during the waiting time after the end of
radiotherapy. This concept has been studied in the recently
closed RAPIDO trial32 and in a recently published Polish
trial.32 Results from the RAPIDO trial are not yet available
but the Polish trial reported improved tolerability and
improved survival after short-course radiotherapy followed
by consolidation chemotherapy compared to conventional
chemoradiotherapy.32 The present aim in rectal cancer
treatment must be to maintain a low rate of local
recurrence, minimise the risk of early and late treatment
toxicity and postoperative complications, and to address
the problem of distant disease. This might be achieved
with short-course radiotherapy with delay and chemo-
therapy in the period between radiotherapy and surgery.
To conclude, short-course radiotherapy with surgery
delayed for 4–8 weeks might have certain advantages over
immediate surgery in rectal cancer treatment. Oncological
outcomes seem similar to short-course radiotherapy with
surgery within a week; acute radiation toxicity is observed
but the postoperative complications are significantly fewer.
Additionally, short-course radiotherapy with delay gives an
opportunity to optimise patients, such as to cease smoking,
initiate an individualised training programme, adjust
blood pressure, provide nutritional support or other

10 www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4


medical interventions, and plan surgery well in advance.
Long-course radiotherapy with delay seems to be no
different than short-course radiotherapy with delay, but
prolongs the treatment time substantially.
Contributors
TH and BC designed and planned the study initially and together with
BG amended the protocol in 1999. AM was the national principal
investigator since 2007 and was responsible for coordinating the trial.
AM, TH, BG, ÅB, DP, MM, FH, CR, OH, and IS enrolled patients and
collected data at the study centres. JE, AM, BG, and TH collected data,
were responsible for registry outtakes, data analyses, and drafting the
report. HJ was responsible for power calculations and statistical
guidance. All authors approved the final version of the report.
Declaration of interests
We declare no competing interests.
Acknowledgments
The study was supported financially by the Swedish Research Council,
the Swedish Cancer Society, and the Stockholm Cancer Society. Financial
support was also provided through the regional agreement on medical
training and clinical research between the Stockholm County Council
and Karolinska Institutet. For the complete list of acknowledgements see
the appendix (p 9).
References
1 Frykholm GJ, Glimelius B, Pahlman L. Preoperative or postoperative
irradiation in adenocarcinoma of the rectum: final treatment results
of a randomized trial and an evaluation of late secondary effects.
Dis Colon Rectum 1993; 36: 564–72.
2 Cedermark B, Johansson H, Rutqvist LE, Wilking N. The Stockholm I
trial of preoperative short term radiotherapy in operable rectal
carcinoma. A prospective randomized trial. Stockholm Colorectal
Cancer Study Group. Cancer 1995; 75: 2269–75.
3 Martling A, Holm T, Johansson H, Rutqvist LE, Cedermark B,
Stockholm Colorectal Cancer Study G. The Stockholm II trial on
preoperative radiotherapy in rectal carcinoma: long-term follow-up
of a population-based study. Cancer 2001; 92: 896–902.
4 Kapiteijn E, Marijnen CA, Nagtegaal ID, et al.
Preoperative radiotherapy combined with total mesorectal excision
for resectable rectal cancer. N Engl J Med 2001; 345: 638–46.
5 Sebag-Montefiore D, Stephens RJ, Steele R, et al. Preoperative
radiotherapy versus selective postoperative chemoradiotherapy in
patients with rectal cancer (MRC CR07 and NCIC-CTG C016): a
multicentre, randomised trial. Lancet 2009; 373: 811–20.
6 No authors listed. Improved survival with preoperative radiotherapy
in resectable rectal cancer. Swedish Rectal Cancer Trial.
N Engl J Med 1997; 336: 980–87.
7 Glimelius B. Optimal time intervals between pre-operative
radiotherapy or chemoradiotherapy and surgery in rectal cancer?
Front Oncol 2014; 4: 50.
8 Pettersson D, Cedermark B, Holm T, et al. Interim analysis of the
Stockholm III trial of preoperative radiotherapy regimens for rectal
cancer. Br J Surg 2010; 97: 580–87.
9 Pettersson D, Lorinc E, Holm T, et al. Tumour regression in the
randomized Stockholm III Trial of radiotherapy regimens for rectal
cancer. Br J Surg 2015; 102: 972–78.
10 Pahlman L, Bohe M, Cedermark B, et al. The Swedish rectal cancer
registry. Br J Surg 2007; 94: 1285–92.
11 Cox JD, Stetz J, Pajak TF. Toxicity criteria of the Radiation Therapy
Oncology Group (RTOG) and the European Organization for
Research and Treatment of Cancer (EORTC).
Int J Radiat Oncol Biol Phys 1995; 31: 1341–46.
12 Rasanen M, Carpelan-Holmstrom M, Mustonen H, Renkonen-
Sinisalo L, Lepisto A. Pattern of rectal cancer recurrence after
curative surgery. Int J Colorectal Dis 2015; 30: 775–85.
13 Kodeda K, Johansson R, Zar N, et al. Time trends, improvements
and national auditing of rectal cancer management over an 18-year
period. Colorectal Dis 2015; 17: O168–79.
14 Blomqvist L, Glimelius B. The ‘good’, the ‘bad’, and the ‘ugly’ rectal
cancers. Acta Oncol 2008; 47: 5–8.
www.thelancet.com/oncology Published online February 9, 2017 http://dx.doi.org/10.1016/S1470-2045(17)30086-4
Articles
15 Sauer R, Becker H, Hohenberger W, et al. Preoperative versus
postoperative chemoradiotherapy for rectal cancer. N Engl J Med
2004; 351: 1731–40.
16 Pach R, Kulig J, Richter P, Gach T, Szura M, Kowalska T.
Randomized clinical trial on preoperative radiotherapy 25 Gy in
rectal cancer—treatment results at 5-year follow-up.
Langenbecks Arch Surg 2012; 397: 801–07.
17 Ngan SY, Burmeister B, Fisher RJ, et al. Randomized trial of
short-course radiotherapy versus long-course chemoradiation
comparing rates of local recurrence in patients with T3 rectal
cancer: Trans-Tasman Radiation Oncology Group trial 01.04.
J Clin Oncol 2012; 30: 3827–33.
18 Bosset JF, Collette L, Calais G, et al. Chemotherapy with preoperative
radiotherapy in rectal cancer. N Engl J Med 2006; 355: 1114–23.
19 Braendengen M, Tveit KM, Berglund A, et al. Randomized phase III
study comparing preoperative radiotherapy with chemoradiotherapy
in nonresectable rectal cancer. J Clin Oncol 2008; 26: 3687–94.
20 Bujko K, Nowacki MP, Nasierowska-Guttmejer A, Michalski W,
Bebenek M, Kryj M. Long-term results of a randomized trial
comparing preoperative short-course radiotherapy with preoperative
conventionally fractionated chemoradiation for rectal cancer.
Br J Surg 2006; 93: 1215–23.
21 Petrelli F, Sgroi G, Sarti E, Barni S. Increasing the interval between
neoadjuvant chemoradiotherapy and surgery in rectal cancer:
A Meta-analysis of Published Studies. Ann Surg 2016; 263: 458–64.
22 Lefevre JH, Mineur L, Kotti S, et al. Effect of interval (7 or 11 weeks)
between neoadjuvant radiochemotherapy and surgery on complete
pathologic response in rectal cancer: a multicenter, randomized,
controlled trial (GRECCAR-6). J Clin Oncol 2016; 34: 3773–80.
23 Hatfield P, Hingorani M, Radhakrishna G, et al. Short-course
radiotherapy, with elective delay prior to surgery, in patients with
unresectable rectal cancer who have poor performance status or
significant co-morbidity. Radiother Oncol 2009; 92: 210–14.
24 Pettersson D, Holm T, Iversen H, Blomqvist L, Glimelius B,
Martling A. Preoperative short-course radiotherapy with delayed
surgery in primary rectal cancer. Br J Surg 2012; 99: 577–83.
25 Radu C, Berglund A, Pahlman L, Glimelius B. Short-course
preoperative radiotherapy with delayed surgery in rectal cancer—
a retrospective study. Radiother Oncol 2008; 87: 343–49.
26 Veenhof AA, Bloemena E, Engel AF, van der Peet DL, Meijer OW,
Cuesta MA. The relationship of histological tumor regression
grade (TRG) and two different time intervals to surgery following
radiation therapy for locally advanced rectal cancer.
Int J Colorectal Dis 2009; 24: 1091–96.
27 Ansari N, Solomon MJ, Fisher RJ, et al. Acute adverse events and
postoperative complications in a randomized trial of preoperative
short-course radiotherapy versus long-course chemoradiotherapy
for T3 adenocarcinoma of the rectum: Trans-Tasman Radiation
Oncology Group Trial (TROG 01.04). Ann Surg 2016; published
online Sept 14. DOI:10.1097/sla.0000000000001987.
28 Tevis SE, Kohlnhofer BM, Stringfield S, et al.
Postoperative complications in patients with rectal cancer are
associated with delays in chemotherapy that lead to worse disease-
free and overall survival. Dis Colon Rectum 2013; 56: 1339–48.
29 Breugom AJ, Swets M, Bosset JF, et al. Adjuvant chemotherapy
after preoperative (chemo)radiotherapy and surgery for patients
with rectal cancer: a systematic review and meta-analysis of
individual patient data. Lancet Oncol 2015; 16: 200–07.
30 Bujko K, Glimelius B, Valentini V, Michalski W, Spalek M.
Postoperative chemotherapy in patients with rectal cancer receiving
preoperative radio(chemo)therapy: a meta-analysis of randomized
trials comparing surgery +/- a fluoropyrimidine and
surgery + a fluoropyrimidine +/- oxaliplatin. Eur J Surg Oncol 2015;
41: 713–23.
31 Nilsson PJ, van Etten B, Hospers GA, et al. Short-course radiotherapy
followed by neo-adjuvant chemotherapy in locally advanced rectal
cancer—the RAPIDO trial. BMC Cancer 2013; 13: 279.
32 Bujko K, Wyrwicz L, Rutkowski A, et al. Long-course oxaliplatin-based
preoperative chemoradiation versus 5 × 5 Gy and consolidation
chemotherapy for cT4 or fixed cT3 rectal cancer: results of a
randomized phase III study. Ann Oncol 2016; 27: 834–42.
11