Organ‐sparing management of luminal gastrointestinal cancers: current and emerging approaches

Abstract

Treatment of luminal gastrointestinal malignancies with curative intent has traditionally required a combination of systemic treatments and complex surgical resections, which carry risks to function and quality of life. Organ‐sparing approaches to treating gastrointestinal malignancies have advanced significantly. With increasing efficacy of upfront systemic and radiation therapies, select patients can safely avoid radical surgery while maintaining oncologic outcomes. This review discusses current and future organ‐sparing approaches in the management of gastro‐oesophageal, colon, rectal and anal cancers.

Introduction

In recent years, neoadjuvant therapies for locally advanced gastrointestinal (GI) cancers have dramatically changed the treatment landscape. With increasing efficacy of upfront systemic and radiation therapies, a portion of patients can achieve a complete clinical response (cCR) prior to curative‐intent surgery. In the treatment of some GI cancers, this raises a key clinical question: in patients where no residual tumour is detected after neoadjuvant therapy, is radical surgery necessary? Here, we explore the mounting evidence for organ‐sparing or non‐operative management of luminal GI cancers.

GI cancers account for approximately 26% of global cancers and 35% of cancer‐related deaths. ¹ Incidence continues to rise, particularly in regions adopting a Western lifestyle characterised by high‐fat diets, obesity and sedentary behaviour. Moreover, high‐income countries have noted an alarming increase in young‐onset GI cancers diagnosed before age 50, without known genetic predisposition, where permanent loss of organ function can have drastic long‐term sequelae. ¹

Surgical resection remains the cornerstone of treatment for most localised luminal GI malignancies. However, major GI surgery carries substantial peri‐operative risks and long‐term morbidity. Complications include haemorrhage, anastomotic leak, infection, respiratory failure and resultant mortality, as well as permanent functional impairment causing incontinence, dysphagia, reflux, dumping syndrome, malnutrition, chronic pain and reduced quality of life (QoL). ² , ³ , ⁴

Organ‐sparing strategies aim to cure or eradicate cancer without surgical resection, thereby preserving the structure and function of the affected organ. A key distinction is that non‐operative management of resectable tumours is dependent on an excellent response to planned neoadjuvant treatment with systemic and/or radiation therapy, with surgery deferred or omitted. In contrast, definitive chemoradiotherapy (CRT) – often reserved for unresectable tumours or patients unfit for surgery – is delivered as a stand‐alone curative approach, involving higher cumulative doses of radiotherapy (RT) without subsequent surgery, followed by surveillance.

Central to organ preservation is achieving cCR, defined as the absence of detectable tumour on clinical examination, endoscopy and imaging (with or without confirmatory biopsy) following neoadjuvant therapy. In contrast, a pathological complete response (pCR) requires histological confirmation of complete tumour regression from a resected specimen.

Therapeutic advances have expanded the feasibility of organ preservation. Definitive CRT has long been established in anal cancer, achieving high rates of local control without surgery. ⁵ More recent evidence supports a ‘watch and wait’ (W&W) approach in some rectal and oesophageal cancers after neoadjuvant CRT (nCRT). ⁶ , ⁷ Squamous cell carcinomas (SCCs), which are exquisitely radiosensitive, and mismatch repair‐deficient (dMMR) or microsatellite instability‐high (MSI‐H) cancers, which can dramatically respond to immunotherapy, are particularly well suited to an organ‐sparing approach (Fig. 1).

Figure 1.

Potential management pathway for curable, locally advanced luminal gastrointestinal cancers stratified by mismatch repair status. Patients with mismatch repair‐proficient tumours undergo total neoadjuvant therapy, followed by multidisciplinary assessment of response. Those with incomplete or partial response proceed to radical surgery with or without chemoradiotherapy and chemotherapy, whereas patients achieving a complete clinical response may enter a watch‐and‐wait strategy with ongoing surveillance. In contrast, mistmatch repair‐deficient tumours are treated with upfront immunotherapy, with subsequent response assessment guiding either watch‐and‐wait management in cases of cCR or surgical intervention with or without chemoradiotherapy if response is incomplete. Local tumour regrowth during surveillance prompts consideration of salvage radical surgery, while sustained cCR supports continued non‐operative management.

When applied appropriately, non‐operative treatments can achieve oncologic outcomes comparable to conventional surgery, ideally while minimising morbidity and maintaining QoL. ³ , ⁴ , ⁸ From a consumer perspective, less invasive options are often favoured; in discrete‐choice experiments, patients report a willingness to accept modestly reduced survival to avoid the physical and psychological burdens of radical surgery. ⁹ However, while some patients with cCR are effectively cured without surgery, the risks of undetected residual disease and subsequent regrowth require rigorous post‐treatment surveillance for early detection and treatment of recurrent disease. Careful patient selection for strict adherence is essential.

Oesophageal cancer

Locally advanced, resectable oesophageal and gastro‐oesophageal junction (GOJ) cancers are conventionally managed with peri‐operative chemotherapy (preferred in adenocarcinoma) or nCRT (preferred in SCC) followed by radical surgical resection – with nearly a quarter of adenocarcinomas and half of SCCs having a pCR at surgery. ⁶ More recent evidence has established adjuvant immunotherapy for patients with residual disease following nCRT and peri‐operative chemo‐immunotherapy. ¹⁰ Despite this progress, the recurrence rate remains high and survival outcomes poor. ¹¹

Although some data suggest that definitive CRT yields similar outcomes in resectable oesophageal SCC, to date, surgical resection with oesophagectomy remains the standard of care for fit patients. However, oesophagectomy is associated with significant perioperative mortality, with reported 30‐day mortality up to 10%. ¹² For select patients with cCR after nCRT, active surveillance may allow deferral or avoidance of surgery. ¹³ An argument against organ‐sparing management of oesophageal cancer is the difficulty of accurately detecting residual disease after neoadjuvant therapy. This was addressed in the preSANO trial, a prospective diagnostic cohort study that assessed the accuracy of cCR evaluations after nCRT for locally advanced oesophageal adenocarcinoma or SCC. ¹⁴ preSANO demonstrated that 90% of substantial residual tumours could be detected using a multimodal strategy combining endosonography with fine‐needle aspiration (FNA) of suspicious lymph nodes, endoscopy with bite‐on‐bite biopsies and positron emission tomography computed tomography. Notably, 31% of substantial residual tumours were missed by endoscopy with regular biopsy and FNA, highlighting the need for rigorous disease assessment when considering non‐operative management.

A 2022 meta‐analysis of seven studies involving 788 patients showed that active surveillance following nCRT yielded overall survival (OS) comparable to immediate surgery in patients with cCR (hazard ratio (HR) 1.08, 95% CI 0.62–1.87, P = 0.75). ⁸ Five‐year locoregional recurrence rate during active surveillance was 40%. Among active surveillance patients who required postponed oesophagectomy for locoregional recurrence, 95% had a complete (R0) resection.

This is supported by the SANO trial (NCT05491616), a multicentre, cluster‐randomised phase III study comparing active surveillance with standard oesophagectomy in 309 Dutch patients who achieved cCR after nCRT for locally advanced oesophageal cancer. ⁶ Participants undergoing active surveillance had clinical response evaluation with imaging and endoscopy every 3 months in year 1, every 4 months in year 2, every 6 months in year 3 and annually in years 4–5 – a highly intensive follow‐up regimen. Early results show non‐inferiority of OS after active surveillance compared to standard surgery (2‐year OS 74% (95% CI 69–78) vs 71% (95% CI 62–78); HR 1.14, 95% CI 0.74–1.78, P = 0.55). Results concerning longer‐term survival are pending. Participants in the active surveillance group – including those who required delayed oesophagectomy for locoregional regrowth – showed significantly better global health‐related QoL at 6‐ and 9‐months compared to those who underwent immediate surgery.

Similarly, the ongoing randomised phase III Canadian NEEDS trial (NCT04460352) is comparing nCRT followed by oesophagectomy (standard therapy) to definitive CRT followed by active surveillance and salvage oesophagectomy if needed for oesophageal SCC. ¹³ It aims to assess non‐inferiority of OS and superiority of QoL with definitive CRT for resectable oesophageal SCC. Results are awaited.

Gastric cancer

Non‐operative management is yet to be established as a potential strategy for mismatch repair‐proficient (pMMR) gastric cancer. The relatively low pCR rates (17%) seen in the recent TOPGEAR trial and lack of survival benefit with nCRT added to peri‐operative chemotherapy in gastric or GOJ adenocarcinoma ¹⁵ emphasise the differing tumour biology of adenocarcinoma compared to SCC and the challenges this poses for organ‐sparing management. In HER2‐positive pMMR gastric adenocarcinoma, the addition of HER2 targeted therapy and immunotherapy to chemotherapy has demonstrated encouraging initial results, with a pCR rate of almost 50% and major pathological response (MPR) rate of 68% in the single‐arm phase II PHERFLOT trial. ¹⁶ Larger confirmatory trials exploring an organ sparing strategy in this population are needed.

In contrast, dMMR gastric and GOJ tumours have shown exceptional response to neoadjuvant immunotherapy in small single‐arm trials (Fig. 1). This falls in stark contrast to peri‐operative chemotherapy in this dMMR population, which has shown little to no benefit above surgery alone, with retrospective analyses demonstrating reduced survival in patients with dMMR disease who received 5‐fluorouracil and oxaliplatin chemotherapy in addition to surgery. ¹⁷

The phase II GERCOR NEONIPIGA study investigated neoadjuvant nivolumab plus ipilimumab followed by adjuvant nivolumab in patients with localised dMMR or MSI‐H gastric or GOJ adenocarcinoma. ¹⁸ The regimen demonstrated remarkable efficacy, with pCR 58.6% (90% CI 41.8–74.1) among patients who proceeded to surgery and nearly all resected tumours showing MPR. Notably, two patients who achieved cCR after neoadjuvant therapy declined surgery and instead received adjuvant nivolumab, highlighting the potential for organ preservation in this population.

Building on this evidence, the phase II INFINITY study (NCT04817826) evaluated neoadjuvant tremelimumab and durvalumab in resectable dMMR/MSI‐H gastric and GOJ adenocarcinomas. ¹⁹ In cohort 1, the activity and safety of tremelimumab plus durvalumab (T300/D) were assessed as neoadjuvant treatment prior to standard surgery. pCR rate was 60% and 24‐month progression‐free survival and OS were 85% and 92%, respectively. Cohort 2 explored the activity and safety of T300/D as definitive treatment in patients achieving cCR, enabling non‐operative management. Of 13 patients undergoing active surveillance, one patient had local regrowth and required salvage surgery to date. These results suggest that biomarker‐directed immunotherapy may allow safe organ preservation in select patients; more evidence and long‐term follow‐up are required.

Colon cancer

Colon cancers present significant challenges to non‐operative management. RT is rarely feasible due to safety concerns and technical limitations in delivery of targeted doses to the colon. ²⁰ There is a risk of neoadjuvant chemotherapy impeding fitness for surgery or increasing perioperative complications, and chemo‐resistant tumours may progress during treatment and compromise the chance of cure. Accurate preoperative staging remains difficult; low‐risk patients may be exposed to chemotherapy when surgery alone may have been sufficient. ²⁰ To date, neoadjuvant chemotherapy has not demonstrated a survival benefit in locally advanced colon cancer and is not used in routine practice. Moreover, ongoing endoscopic and radiological surveillance required for non‐operative management is complex, resource‐intensive, and burdensome for patients, making timely detection of residual or recurrent disease difficult. ²⁰ Consequently, upfront surgery remains the standard of care for operable, MMR‐proficient colon cancer. ²¹

In contrast, recent advancements in neoadjuvant immunotherapy have shown promising results for early‐stage dMMR colon tumours, which comprise up to 15% of non‐metastatic colon cancers. ²² The phase I NICHE trial evaluated neoadjuvant nivolumab and ipilimumab in early‐stage dMMR colon cancer. ²¹ All 20 patients included in the study exhibited a pathological response, and 60% achieved pCR. These results were confirmed in NICHE‐2, a larger multicentre phase II trial of 112 patients with dMMR colon cancer which reported 67% pCR and 95% MPR, with no recurrences at 36‐month follow‐up. ²² , ²³ A systematic review by Chakrabarti et al. analysed 37 studies involving 423 patients with dMMR colon (77%) and rectal (23%) early‐stage cancers (88% stage III) and reported that neoadjuvant immunotherapy was associated with high CR rate (pCR + cCR; 72% (305 of 423)) and excellent safety profile. ²⁴

Several phase II trials are now evaluating an active surveillance or W&W strategy in locally advanced MSI‐H/dMMR colorectal cancer. In the Chinese NEOCAP trial of neoadjuvant camrelizumab plus apatinib in this population, 54% of patients had cCR and were managed with W&W; at median follow‐up of 11.2 months, no regrowth has been observed. ²⁵ The WINDOW trial is assessing serial ctDNA as an approach for determining cCR in this population. ²⁶ Of 24 included patients, 22 achieved ctDNA negativity during upfront immunotherapy and proceeded with active surveillance; thus far, none has recurred. The randomised phase II PREMICES trial is currently evaluating a non‐surgical approach in dMMR resectable colorectal cancer patients. ²⁷ The experimental arm is neoadjuvant pembrolizumab followed by a W&W strategy with a novel primary endpoint of ‘6‐month strategy success’ defined as the absence of death, disease progression or the need for primary tumour resection during that period.

Data to date show the potential of immunotherapy alone to achieve disease control in dMMR/MSI‐H early‐stage or non‐metastatic colon cancer. Continued evaluation in prospective trials with validation of surveillance strategies is required before integration into standard clinical practice.

Rectal cancer

In the last decade, total neoadjuvant therapy (TNT) has redefined the management of locally advanced rectal cancer (LARC) by administering both CRT and chemotherapy entirely before surgery. TNT increases tumour regression, treatment compliance and pCR, making a W&W approach possible for selected patients who achieve cCR. ²⁸ These patients can defer or avoid surgery, reducing morbidity while maintaining favourable oncologic outcomes. Up to 30% develop local regrowth within 3 years, so careful selection and strict surveillance are essential, with salvage surgery feasible in most cases. ³ , ⁴ , ²⁸ , ²⁹

The W&W approach in LARC was pioneered by Habr‐Gama, who evaluated tumour response after nCRT and enrolled patients with cCR in a rigorous surveillance programme involving monthly clinical assessments and imaging (abdominal and pelvic CT scans and chest radiographs). ²⁸ Patients showing incomplete or partial response were referred for radical surgery. In their landmark 2004 report, 71 patients (26.8%) achieved cCR 8 weeks after completing CRT and were enrolled in the observation group. ²⁸ After a mean follow‐up of 57.3 months, the overall recurrence rate was 7%, with no cancer‐related deaths. Five‐ and 10‐year overall and disease‐free survival rates were similar in the observation and resected groups.

Subsequent prospective studies formalised a ‘W&W’ policy. Maas et al. showed that deferring surgery following cCR is both feasible and safe. ⁴ Disease‐free and OS were comparable between wait‐and‐see patients and those undergoing surgery (2‐year disease‐free survival 89% (95% CI 43–98) vs 93% (95% CI 59–99), P = 0.770; 2‐year OS 100% vs 91% (95% CI 59–99), P = 0.228). Patients in the wait‐and‐see group self‐reported significantly improved bowel function compared to the control group.

The International Watch & Wait Database (IWWD) registry study later reinforced these findings. ³⁰ In 880 patients with cCR, local regrowth occurred in 213 patients (2‐year actuarial rate of 25.3%). The vast majority (88%) of local regrowth occurred within the first 2 years. Of 213 patients with regrowth, 148 had details of surgery available; 99% proceeded to surgery with curative intent, with 88% of salvage resections achieving clear margins. The study reported a 5‐year OS rate of 85% (95% CI 80.9–87.7) and a 5‐year disease‐specific survival of 94% (95% CI 90.8–95.8).

The randomised phase II OPRA trial prospectively evaluated selective non‐operative management in stage II–III rectal cancer, assessing induction chemotherapy followed by CRT (INCT‐CRT) and CRT followed by consolidation chemotherapy (CRT‐CNCT) compared to historical controls. ³¹ Of 304 patients (74%), 225 achieved cCR or near‐cCR and were offered W&W. Tumour regrowth occurred in 40% of the INCT‐CRT group and 27% of the CRT‐CNCT group. At 3 years, rectal preservation was achieved in 41% in the INCT‐CRT group and 53% of patients in the CRT‐CNCT. Distant metastasis‐free survival was similar in both groups (84% INCT and 82% CNCT). These findings support a hypothesis of improved local control with upfront CRT followed by consolidation chemotherapy, but this has not been confirmed by phase III data.

The Australian RENO trial (ACTRN12619000207112) is prospectively assessing adult patients with rectal cancer treated with CRT‐CNCT strategy. ³² Patients who achieve cCR are managed with a W&W approach with strict surveillance (clinical review, magnetic resonance imaging rectum and flexible sigmoidoscopy 3‐monthly for 2 years, then 6‐monthly for 3 years) and salvage surgery in the event of local regrowth. Patients with residual disease proceed to standard surgical resection. The co‐primary outcomes are the 3‐year local failure rate and rate of rectal preservation, with secondary outcomes including safety, QoL and patient‐reported outcomes. Results are awaited.

As in other dMMR cancers, evidence supports upfront immunotherapy for dMMR LARC. A phase II study evaluated a 6‐month course of dostarlimab, a PD‐1‐blocking monoclonal antibody, in 50 patients with locally advanced dMMR rectal cancer. Among 49 patients who completed therapy, all achieved cCR, with no patient requiring chemotherapy, radiation or surgery. ³³ Updated results in 2025 reported one local recurrence (that was not resected), with a 96% recurrence‐free survival at 2 years. ³⁴ Three of the 50 patients were subsequently able to conceive and deliver healthy babies, which might not have been possible with conventional treatment.

Despite this growing body of evidence supporting W&W for LARC with cCR after TNT, there is ongoing equipoise regarding optimal management – the benefits of organ preservation must be carefully weighed against the need for strict surveillance and potential for local regrowth and metastatic disease. W&W has been endorsed in international guidelines where organ preservation is desired (early‐stage disease, patient preference) and cCR is achieved. ³⁵ However, a recent retrospective analysis of 793 patients included in the IWWD reported local regrowth at any time as an independent risk factor for metastatic disease (regardless of successful salvage surgery). ³⁶ Overall, 85 patients (10%) developed metastatic disease – a rate similar to patients undergoing surgery. However, of these, 51 (60%) had experienced local regrowth at any time, highlighting the need for careful counselling for patients considering W&W and further research to identify patients who are at high risk of local regrowth, which may confer a higher risk of subsequent metastatic disease. Data are limited regarding peri‐operative risk of deferred or salvage surgery compared to standard of care surgery. One study analysing peri‐operative outcomes in 78 W&W patients (of whom 23 underwent surgery) reported similar rates of anastomotic leaks, 30‐day morbidity, reintervention and readmission rates and no difference in pathology features or 3‐year oncological outcomes. ³⁷

While modern TNT protocols combined with strict W&W surveillance are transforming rectal cancer management, careful patient selection and protocol adherence are critical, and this burden is not to be underestimated.

Anal cancer

In the 1970s and 1980s, Nigro and colleagues demonstrated that combining 5‐fluorouracil and mitomycin C (5‐FU/MMC) with radiation produced complete responses in many patients with anal SCC. This regimen eliminated the need for upfront surgery and established the basis for organ‐sparing approaches in anal cancer. ⁵

Building on this foundation, the EORTC 1997 trial randomised 110 patients with locally advanced anal cancer (T3–4 or N1–3) to RT alone or RT combined with 5‐FU/MMC. ³⁸ While OS was similar between groups, CRT produced a significantly higher pathologic remission rate (80% vs 61%; P = 0.02). By avoiding abdominoperineal resection (APR), which removes the anal sphincter and requires a permanent colostomy, patients can maintain continence and experience a substantially improved QoL.

The RTOG 98‐11 trial compared concurrent chemoradiation with 5‐FU/MMC versus 5‐FU plus cisplatin in anal cancer. Long‐term results showed that while OS was similar, the MMC‐based regimen significantly reduced colostomy rates (10% vs 19%; P = 0.02), demonstrating superior organ preservation. ³⁹ The ACT II trial confirmed that replacing MMC with cisplatin offered no improvement in disease‐free or OS, reaffirming MMC as the preferred agent. ⁴⁰

Despite these advances, CRT fails in 20%–30% of patients, with persistent disease in 10%–15% and local recurrence in 10%–15%. For these patients, high‐morbidity salvage APR, often with myocutaneous flap reconstruction, is the only option to achieve local control. Vigilant follow‐up is mandatory. Further research is under way to improve outcomes by optimising RT and incorporating novel systemic agents.

Patient perspectives on non‐operative management

Long‐term toxicity and functional preservation are increasingly important when weighing the benefits of organ‐sparing strategies. While radical surgery carries a well‐recognised burden of long‐term morbidity, non‐operative management offers the possibility of avoiding these sequelae in appropriately selected patients. Custers et al. demonstrated that W&W patients report favourable long‐term QoL, including lower rates of major low anterior resection syndrome, better social functioning and greater overall functional stability compared with postoperative cohorts. A comprehensive review similarly emphasises that patient‐reported outcome measures consistently favour organ preservation, highlighting better day‐to‐day functioning, improved psychosocial outcomes and greater satisfaction with treatment. ⁴¹ These data reinforce the potential for non‐operative management to improve long‐term QoL and patient‐centred outcomes.

Conclusion

Organ preservation in luminal GI cancers has advanced significantly, with evidence supporting the selection of patients who can safely avoid radical surgery while maintaining oncologic outcomes. In oesophageal cancer, findings from the SANO trials suggest that active surveillance after CRT can safely spare select patients from oesophagectomy. ⁶ In rectal cancer, the OPRA trial and data from the IWWD dataset have demonstrated that patients achieving cCR after TNT could defer surgery, achieving organ preservation in roughly half of cases without compromising disease‐free survival. ³⁰ , ³¹ The ongoing Australian RENO trial seeks to standardise surveillance and further validate non‐operative management. ³² In anal cancer, the EORTC 1997, RTOG 98‐11 and ACT II trials found that 5‐FU/MMC remains the standard chemotherapeutic approach for sphincter preservation. ³⁸ , ³⁹ , ⁴⁰ The impressive rates of pCR following neoadjuvant immunotherapy in dMMR tumours across the GI spectrum highlight the potential for organ preservation in these patients, with further outcome data awaited. QoL analyses consistently support non‐operative management as superior to radical surgery, and patient preference to avoid surgical morbidity is a key factor in decision‐making.

Despite these promising developments, robust prospective evidence and a more nuanced understanding of patient selection are required before organ‐sparing treatment is widely accepted as standard of care. Looking ahead, the challenge is to refine organ‐preserving strategies by identifying those patients who can safely avoid upfront surgery without compromising long‐term outcomes. Translational research defining the spectrum of ‘cCR’, increasing use of ctDNA as a marker of minimal residual disease and novel therapeutic techniques, including stereotactic and proton RT, is likely to expand the safety and clinical utility of non‐operative approaches. Furthermore, trials that test novel combinations of immunotherapy, targeted therapies, chemotherapy and radiation therapy may expand the pool of candidates for organ preservation.

Organ preservation, however, comes with the trade‐off of rigorous surveillance and the potential need for delayed surgery. Beyond treatment efficacy, there is a pressing need to standardise surveillance protocols, optimise salvage strategies and understand the psychosocial and functional outcomes of non‐operative care. Ultimately, the future of organ preservation in GI cancers will depend on integrating these innovations into a personalised framework that balances oncologic safety with QoL.

Data availability statement

Data sharing was not applicable to this article as no datasets were generated or analysed during the current study.