Is “watch-and-wait” after chemoradiotherapy safe in patients with rectal cancer?
BMJ 2018; 363 doi: https://doi.org/10.1136/bmj.k4472 (Published 13 November 2018) Cite this as: BMJ 2018;363:k4472
- Fraser M Smith, consultant general/colorectal surgeon, and chair of Cheshire and Mersey Contact Therapy/Watch and Wait MDT1 4,
- Katharine Cresswell, patient and public involvement and engagement project manager2,
- Arthur Sun Myint, consultant clinical oncologist and lead clinician in the Papillon Suite, and honorary professor, University of Liverpool3 4,
- Andrew G Renehan, professor of cancer studies and surgery, and consultant in colorectal and peritoneal surgery5 6
- 1Department of Surgery, Royal Liverpool and Broadgreen University Hospitals NHS Trust, Liverpool, UK
- 2NIHR Biomedical Research Centre Patient and Public Cancer Research Advisory Group, Public Programmes Team, Manchester University NHS Foundation Trust; and Faculty of Biology, Medicine and Health, University of Manchester, Manchester, UK
- 3Clatterbridge Cancer Centre, Bebington, Wirral, UK
- 4Institute of Translational Medicine, University of Liverpool, Liverpool, UK
- 5Manchester Cancer Research Centre, NIHR Manchester Biomedical Research Centre, Division of Cancer Sciences, School of Medical Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Manchester, UK
- 6Colorectal and Peritoneal Oncology Centre, Christie NHS Foundation Trust, Manchester, UK
- Correspondence to: F M Smith fraser.smith{at}rlbuht.nhs.uk
What you need to know
After standard long-course chemoradiotherapy for locally advanced rectal cancer, up to a quarter of patients have no clinically apparent tumour—referred to as a clinical complete response
Evidence from observational studies suggests these patients can be considered for a “watch-and-wait” approach with regular surveillance to avoid major surgery
Up to a third of patients on a watch-and-wait programme develop tumour regrowth and require salvage surgery; the long term outcomes are uncertain
Colorectal cancer is the third commonest cancer worldwide.1 About a third of these arise in the rectum. Approximately a third of rectal cancers are locally advanced and at high risk of recurrence. Long-course chemoradiotherapy followed by surgical resection is now standard treatment for these tumours in the UK,23 Europe,4 the US,5 and Australia.6 However, surgery is associated with major complications (up to 15%), perioperative mortality (up to 5%), and the need for a permanent stoma in up to a quarter of patients.7
Within the published literature, after chemoradiotherapy 10-25% of patients have no residual tumour on pathological examination after surgical resection.78910 Clinical examination before surgery in these patients has shown an equivalent favourable response, referred to as a clinical complete response. The criteria for defining a clinical complete response include absence of residual ulceration, stenosis, or mass within the rectum during digital rectal examination and endoscopic examination (fig 1).910
Endoscopic images of rectum after chemoradiotherapy for rectal cancer show (A) “white” scar (circled) and (B) telangiectasia changes (arrow). Both are accepted as “normal” and are compatible with a clinical complete response
Nearly 15 years ago, a seminal study from a centre in São Paulo, Brazil reported that 71 patients with clinical complete response were managed without initial surgery. Instead they were managed by a non-surgical “watch-and-wait” approach and had equivalent survival compared with patients with complete response managed by major surgery.11 Several centres have since replicated these results, but mainly in small case series.
In a “watch-and-wait” approach, patients may opt to forgo surgery, typically 8-12 weeks after the end of long-course chemoradiotherapy, and instead enter a surveillance programme of monitoring with regular digital and endoscopic examination supplemented with magnetic resonance imaging (box 1). There is uncertainty about the safety of watch-and-wait, with concerns that (in the absence of surgical resection) subclinical residual cancer cells may be undertreated and re-manifest later as inoperable local recurrence or as metastatic disease and compromise survival.12
Example of a “watch-and-wait” approach to rectal cancer after clinical complete response to chemoradiotherapy
Years 1-2
The most intensive monitoring period after chemoradiotherapy
Patients will see a named organ preservation surveillance specialist every 3 months and undergo a digital rectal examination and flexible sigmoidoscopy. This should aim to incorporate narrow band imaging.
Pelvic magnetic resonance imaging (MRI) will be carried out every 4-6 months
Carcinoembryonic antigen (CEA) blood test every 4 months
Computed tomography (CT) of thorax, abdomen, and pelvis performed every 6 months
Colonoscopy carried out as per NICE surveillance guidelines
Years 3-5
If clinical complete response is maintained, the frequency of flexible sigmoidoscopy and digital rectal examination will be reduced to every 6 months in year 3 and yearly thereafter up to 5 years
CEA blood test every 6 months
Patient’s last MRI at 36 months, as relapse rate is very low after this
One further CT of thorax, abdomen, and pelvis at 36 months
After 5 years
Annual digital rectal examination, flexible sigmoidoscopy, and CEA blood test
MRI or CT based on clinical suspicion
Suspicion or confirmation of tumour regrowth at any stage in surveillance programme
Patient should be referred to colorectal surgeon immediately for full work-up for radical surgery
What is the evidence of uncertainty?
Our literature search (box 2) identified three meta-analyses of observational data, two of study-level data1516 and one of individual participant data17; a matched-controlled study18; and an international registry, including published and unpublished data19 (see table 1). We found no published randomised controlled trial. Overall survival and disease-free survival are considered the metrics of oncological safety (box 3). The present evidence suggests that, after long-course chemoradiotherapy, patients on watch-and-wait have similar overall and disease-free survival compared with patients who undergo surgical resection. Around 30% of patients initially thought to have a complete response and opting for watch-and-wait experience tumour regrowth, most commonly in the first two years. Approximately 85% of patients with local regrowth are suitable for surgery and have no oncological disadvantage compared with patients who undergo immediate surgery. Most studies report on outcomes at two to three years, and longer term evidence is lacking. The International Watch and Wait Database (IWWD),19 the largest series of patients with rectal cancer managed by watch-and-wait (880 patients), noted that five-year overall survival was 85% (95% confidence interval 80.9% to 87.7%) and five-year disease-specific survival was 94% (91% to 96%).
Data sources and selection criteria
We searched for studies with:
Population: Patients with rectal cancer treated by long-course chemoradiotherapy who achieved a clinical complete response
Intervention: Watch-and-wait programme for follow-up
Comparator: Surgical resection
Outcome: Local cancer regrowth and patient survival. We do not report on “near complete response,”13 which might be treated by local radiotherapy (such as the Papillon technique)14 or local excision.
We searched PubMed from 1 January 2000 to 31 August 2018 for randomised controlled trials and systematic reviews using the terms “rectal cancer,” “neoadjuvant chemoradiotherapy,” “watch and wait,” “organ preservation,” “randomised controlled trial,” “systematic review” and “meta-analysis.”
We searched clinicaltrials.gov and the PROSPERO registry, and the Cochrane Library to identify ongoing trials and systematic reviews, respectively.
We also reviewed clinical guidelines from the UK National Institute for Health and Care Excellence (NICE 2012), the Association of Coloproctology of Great Britain and Ireland (ACPGBI 2017), European Society of Medical Oncology (ESMO), American Society of Colon and Rectal Surgeons (ASCRS), and the Australasian Cancer Council.
RETURN TO TEXTSummary characteristics and key outcomes from meta-analyses, matched-controlled study, and registry-based pooled analysis
Main outcomes of interest in follow-up after treatment of rectal cancer
Local regrowth—Re-emergence of tumour at the original site during follow up, where it was clinically deemed to have completely regressed. In patients on “watch and wait” pathways this is amenable to subsequent (salvage) surgery with a high chance of cure. This is considered different to local pelvic recurrence in patients who have already had rectal cancer surgery, and who have limited opportunities for salvage.
Overall survival—The time from start of treatment and remaining alive. Our patient liaison group identified overall survival as the key most important factor in their decision making.
Disease-free survival—The time after primary treatment to the re-emergence of any signs (clinical or imaging) or symptoms of that cancer, or death.
There are several limitations of the evidence. First, the absence of randomised controlled trials raises the concern of treatment selection bias. This is partly but not totally accounted for using methodologies such as propensity score matching.18
Second, there is between-study heterogeneity for key predictors of outcomes such as age, performance status, and cancer stage. The meta-analysis of individual participant data illustrates this and shows that some of the variability in local regrowth rates is due to different tumour characteristics, in particular T stage (tumour extent).17 A range of criteria are used to describe indications for long-course chemoradiotherapy across studies. Pre-treatment T stage (determined using magnetic resonance imaging) is a predictor for local regrowth rates. The definition of locally advanced rectal cancer is not internationally standardised, resulting in variations in patient selection for watch-and-wait and in outcomes.
Third, variations in centre-level expertise, treatment, and follow-up protocols might be a source of bias. However, methods to account for these shared centre-level attributes indicate that these do not materially affect local regrowth rates.17
Finally, patient choice can influence outcomes and is difficult to take into account in analyses.
A systematic review and meta-analysis (23 published and unpublished studies, 867 patients) evaluated outcomes in patients with locally advanced rectal cancer managed by a watch-and-wait approach after complete clinical response to long-course chemoradiotherapy.16 The local regrowth rate after two years was 15.7%, with notable heterogeneity between studies and rates ranging from 3.3% to 33%. There was no difference in overall survival compared with patients treated with surgery, but disease-free survival was better in the surgery group. This might reflect differences in study definitions of disease-free survival which are not captured with uniform criteria.21
A second meta-analysis reported similar findings but included published studies only (17 studies, 692 patients); the three-year local regrowth rate was 22% (95% confidence interval 14.3% to 31.8%).15
A recent meta-analysis of individual participant data (11 datasets, 602 patients), the InterCoRe consortium,17 demonstrates that some of the between-study heterogeneity is explained by case mix, such as different pre-treatment T stage.17
The OnCoRe project18 was a “real world” evaluation of the oncological safety of watch-and-wait after clinical complete response across four regions in the North West of England and Wales. In a propensity-score matched cohort analysis (109 watch-and-wait versus 109 resection surgery, matched on age, performance status, and cT stage of cancer), no difference was observed in overall survival and disease-free survival (excluding local regrowth) at three years. A third of patients on watch-and-wait (41 patients) had local regrowth, and 36 (88%) of these were salvaged.
A further systematic review (9 studies, 370 patients) assessed the rate of salvage surgery in patients who developed tumour regrowth on watch-and-wait.20 Of the 114 patients with tumour local regrowth, salvage surgery was possible in 83.8%. The analysis reported no difference in overall survival between patients who received immediate surgery versus the watch-and-wait group.
In the International Watch and Wait Database (IWWD)19 (880 patients with a clinical complete response on watch-and-wait) the two-year cumulative incidence of local regrowth was 25.2% (95% CI 22.2% to 28.5%). There was no comparator. Most (88%) of the local regrowths were in the first two years. With a median of 3.3 years follow-up, distant metastases were diagnosed in 8%, five-year overall survival was 85%, and five-year disease-specific survival was 94%.
Is ongoing research likely to provide relevant evidence?
We searched clinicaltrials.gov to identify ongoing trials and found 16 registered proposals to the search “watch and wait” plus “rectal cancer.” However, against our aforementioned PICO (box 2), we found only one protocol designed to evaluate equivalence in oncological outcomes (three-year disease-free survival) for patients with locally advanced rectal cancer with clinical complete response after long-course chemoradiotherapy randomised to upfront radical surgery or to watch-and-wait (NCT02052921).22 This is a phase II trial from São Paulo, Brazil, which is due to close in 2019. This will be the first randomised trial, and in theory, it should directly address the current uncertainty. However, with a target of 150 patients, this trial seems underpowered, and, in the UK at least, its inclusion criteria will include some tumours at an earlier stage than would normally be considered for long-course chemoradiotherapy.2 The remaining trial protocols include watch-and-wait as part of a component of managements, intended to achieve the goal of “organ preservation.” Nonetheless, trial environments will be important to standardise entry criteria—notably in terms of T stage of tumour.
There is now a debate whether the required safety evidence should come from “hard-to-do” randomised trials or from “scaled-up” analyses of observational data, in order for the watch-and-wait approach to become standard care. We have previously argued that, ideally, a watch-and-wait policy after clinical complete response could be tested against surgery in a randomised controlled trial assessing both oncological and functional results.18 Our clinical experience is that many patients who have a clinical complete response express a strong preference to avoid major surgery, even when explicitly informed about the experimental nature of the watch-and-wait approach.18 This is echoed by other commentators2324 and our Patient and Public Involvement and Engagement (PPIE) team involved in preparing this article (see box on patient involvement). Thus, there is a need for scaled-up analysis of observational data on patient preferences.
What should we do in the light of the uncertainty?
It has not been established which patients might benefit from a watch-and-wait approach. Many patients prefer avoiding surgery. Others may become anxious during watch-and-wait follow-up and may prefer initial surgery as reassurance that their tumour has been removed. In the UK, recent changes in consent law, based on the Montgomery Ruling, stipulate that patients must be informed of any risk that “a reasonable person in the patient’s position would be likely to attach significance to.”19 This rationale, we feel, should not just be limited to the UK as it embodies the concept of patient centred care. We therefore recommend that, to fulfil the mandate of informed consent, patients undertaking long-course chemoradiotherapy are informed that there is a 10-25% chance they will have no residual cancer after the therapy, and, in turn, may consider a watch-and-wait programme with potential avoidance of major surgery. As surgery for rectal cancer includes risks of complications and perioperative mortality, and many patients might require a stoma, most patients are likely to “attach significance to” this information. Recent guidelines (2017) from the UK Association of Coloproctology of Great Britain and Ireland (ACPGBI)3 and Australian Cancer Council6 include watch-and-wait as an option, with the caveat that patients must be informed that it remains a new management under evaluation. The European Society of Medical Oncology (ESMO)4 2018 guidelines state that watch-and-wait can be considered in “high risk” patients (without defining criteria) with clinical complete response after long-course chemoradiotherapy. Older guidelines, such as those from the US (2013),5 recommend that resection is standard care, whereas the UK National Institute for Health and Care Excellence (NICE, 2012)2 does not mention clinical complete response. While we await wider endorsement in clinical guidelines, we recommend a practical guidance for clinicians on how to incorporate watch-and-wait into their practice in box 4 and provide a template watch-and-wait surveillance schedule in box 1.
Practical guidance to clinicians in light of present evidence
Before starting long-course chemoradiotherapy for rectal cancer, discuss with the patient that there is a 10-25% chance of no residual cancer after therapy. This should be included in the patient information sheet.
At 8-12 weeks after long-course chemoradiotherapy, offer restaging to look for a clinical complete response. This constitutes a flexible sigmoidoscopy and digital rectal examination by a clinician experienced in the identification of a clinical complete response (typically a colorectal surgeon), and magnetic resonance imaging.
If a clinical complete response is identified, the colorectal surgeon should discuss the options of resection surgery versus a watch-and-wait approach with the patient.
Offer patients a counselling session with a stoma therapist before making a decision as to whether to undergo surgery if living without a stoma is their main reason for preferring watch-and-wait.
If the patient chooses a watch-and-wait approach, offer monitoring in accordance with a pre-agreed surveillance protocol.
Information on patient and tumour characteristics, treatment, and outcome should be collected and submitted to ongoing registries (such as IWWD) or research databases (such as OnCoRe).
What patients need to know
Surgical removal of the tumour after chemoradiotherapy is the current standard treatment for rectal cancer and is regarded as the most reliable oncological treatment.
Surgery is associated with side effects such as a temporary or permanent stoma bag, pelvic organ dysfunction, sexual dysfunction, bladder and bowel dysfunction, abdominal wall herniation, and a risk of death.
10-25% of rectal cancers can disappear completely after receiving neoadjuvant chemoradiotherapy, meaning that surgery can potentially be avoided in these cases.
Clinical assessment of complete tumour disappearance (“clinical complete response”) can be used to identify patients in whom the tumour is likely to have disappeared on pathological assessment and potentially select them for a non-operative “watch-and-wait” programme. In this case, surgery is reserved only for patients who experience tumour regrowth (see box 1).
To date there is no randomised controlled trial comparing outcomes of watch-and-wait versus radical surgery for patients with clinical complete response, so we are uncertain whether “watch-and-wait” is safe.
“Watch-and-wait” is not currently recognised as a treatment option for healthy patients otherwise suitable for surgery by many formal national guideline bodies.
Around 30% of patients initially thought to have a complete response and opting for watch-and-wait will experience tumour regrowth that requires surgery.
If tumour does regrow, about 85% of patients will be suitable for curative-intent salvage surgery, and they should be at no oncological disadvantage compared with patients who undergo immediate surgery.
Discuss with your surgeon the options of “watch-and-wait” and surgery to determine the appropriate option for you after chemoradiotherapy for rectal cancer.
If you choose “watch-and-wait” you will need frequent rectal examinations, endoscopic procedures, and scans to check that the tumour is not growing back (see box 1).
A patient information sheet is available: http://www.complete-response.com
Education into practice
If you identified a patient with rectal cancer who required long-course chemoradiotherapy, how would you discuss the possibility of a clinical complete response and management options?
Would you register patients receiving long-course chemoradiotherapy for rectal cancer in a national research database?
How patients were involved in this article
We discussed this topic with a sub-group of the NIHR Manchester Biomedical Research Centre (BRC) Patient and Public Cancer Research Advisory Group, facilitated by the Public Programmes Team and the OnCoRe research database coordinator, Mr Lee Malcomson.
The group emphasised that, for the great majority of patients with cancer, the primary outcome of importance is survival. They welcomed the innovation of a “watch-and-wait” approach for patients with rectal cancer as a potential way to avoid major surgery, but sought reassurance from cancer doctors that this “lesser” treatment would not compromise their chances of survival.
Patients emphasised the need to discuss this new treatment pathway carefully and with sensitivity.
Patients asked about alternate options. The literature includes local therapies, such as local excision and radiotherapy, but these have been highly selective reports, and we did not include them in this article.
Acknowledgments
We thank Sharon Williams, a colorectal cancer patient/lay representative, whose comments and input helped shape the paper from a patient’s perspective. We thank Dr Caroline O’Callaghan, a general practitioner, whose comments helped us make the paper accessible to a generalist readership.
Footnotes
This is one of a series of occasional articles that highlight areas of practice where management lacks convincing supporting evidence. The series advisers are Sera Tort, clinical editor, and David Tovey, editor in chief, the Cochrane Library. To suggest a topic for this series, please email us at uncertainties@bmj.com
Funding: This research was supported by the NIHR Manchester Biomedical Research Centre.
Competing interests: The BMJ judged that there were no relevant financial interests. The authors declare: AGR has received lecture honoraria from Merck Serona and Janssen-Cilag, and independent research funding and lecture honoraria from Novo Nordisk (regarding insulins and cancer risk) and Sanofi Pasteur MSD (regarding HPV vaccinations and anal cancer).
Provenance and peer review: Commissioned; externally peer reviewed.