Challenges presented by complete response to immune checkpoint blockade in patients with dMMR colorectal cancer: A case report

Abstract

Introduction

Early clinical trials have demonstrated remarkable responses to immune checkpoint blockade (ICB) in patients with colorectal cancers with deficient mismatch repair (dMMR) mechanisms. The precise role immunotherapy will play in the treatment of these patients is undefined, with these agents likely to produce new challenges as well as opportunities.

Presentation of case

A 74-year-old patient was diagnosed with a locally advanced dMMR adenocarcinoma in the transverse colon with clinical suspicion of peritoneal metastases (cT4N2M1). The burden of disease was assessed as incurable, and a referral was made for palliative oncological treatment. After 5 months of treatment with pembrolizumab, a complete radiological response in the primary tumour was seen although there was still radiological suspicion of peritoneal and lymph node metastases. The patient underwent cytoreductive surgery and hyperthermic intraperitoneal chemotherapy but unfortunately died 6 weeks later due to complications. Final histology of the surgical specimen showed no evidence of residual disease (ypT0N0M0).

Discussion

This case highlights the opportunities and challenges presented by the efficacy of ICB in dMMR colorectal cancer. These agents were able to cure a patient who had disseminated disease presumed to be incurable at the time of diagnosis. However, due to current limitations in determining the degree of response to ICB, this result could only be confirmed after major surgery, which ultimately led to the patient's death.

Conclusion

ICB can lead to dramatic responses in patients with dMMR colorectal cancers. Major challenges remain in differentiating complete and partial responders and determining the indications for conventional surgery.

Highlights

• •Immune checkpoint blockade is effective in patients with dMMR colorectal cancer.
• •The exact role of these agents as neoadjuvant therapies remains to be defined.
• •Identifying complete responders and defining the need for surgery are major challenges.

1. Introduction

The advent of immunotherapy in the form of immune checkpoint blockade (ICB) has led to dramatic improvements in the treatment of several cancers with historically poor prognoses. Arguably the best responses have been seen in patients with metastatic melanoma, where the efficacy of ICB continues to revolutionize the treatment paradigm of these patients [1], [2].

Recent studies have demonstrated the potential of ICB in a subset of patients with colorectal cancer that have deficient DNA mismatch repair mechanisms, otherwise referred to as dMMR colorectal cancers. Early clinical studies have shown marked responses to neoadjuvant ICB in these patients, with overall response rates of up to 100 % and complete pathological responses seen in up to 60 % of patients [3], [4].

Although still to be confirmed in larger studies, these early results have rightly led to optimism regarding the use of ICB in these patients. However, the precise role that immunotherapy will come to play in the management of patients with colorectal cancer remains to be defined, with the efficacy of these agents likely to pose significant challenges in terms of decision making for patients and clinicians alike.

Here we present a case of a patient with a dMMR adenocarcinoma of the colon treated with ICB that highlights some of these challenges. This case is reported in line with SCARE criteria [5].

2. Presentation of case

A 74-year old patient was referred with weight loss, loss of appetite and a palpable abdominal mass. Subsequent endoscopy demonstrated a presumably malignant stenosis in the transverse colon, with biopsies confirming a dMMR colorectal adenocarcinoma. Staging computed tomography (CT) demonstrated local invasion of the abdominal wall in addition to retroperitoneal lymph node metastases and peritoneal metastases in keeping with cT4N2M1 disease. Following multidisciplinary team discussion, the patient's burden of disease was assessed as incurable and they were referred for palliative oncological therapy.

The patient was treated with pembrolizumab at a dose of 2 mg/kg every 3 weeks. No other immunotherapies or cytotoxic chemotherapeutic agents were used. After 4 cycles, the patient was admitted with bowel obstruction and underwent emergency surgery in the form of a defunctioning loop ileostomy. After an uncomplicated recovery, the patient received a further cycle of pembrolizumab before a follow-up CT scan was performed, 5 months after the start of treatment. That CT showed a complete radiological response in the primary tumour (Fig. 1), which was no longer identifiable, although there was still radiological suspicion of peritoneal and lymph node metastases (Fig. 2).

Fig. 1.

Computed tomography images of the primary tumour (green arrow) at diagnosis (A) and following treatment with pembrolizumab (B). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2.

Computed tomography images of pathological lymph nodes (green arrows) at diagnosis (A) and following treatment with pembrolizumab (B). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

On that basis the patient was referred for cytoreductive surgery, including an extended right hemicolectomy, greater omentectomy and bilateral salpingo-oophorectomy followed by hyperthermic intraperitoneal chemotherapy. This consisted of a 90-minute perfusion with 35 mg/m² mitomycin. The patient unfortunately developed an anastomotic leak 3 weeks post-operatively, which was managed with a resection of the anastomosis and formation of an end-ileostomy. This was further complicated by subsequent abdominal dehiscence. The patient had suffered from severe abdominal pain since their initial surgery and, after 6 weeks of admission, decided that they did not want any further active treatment. The patient was discharged to a hospice and died shortly thereafter. The final histological report from the patient's initial surgery demonstrated a complete pathological response with no evidence of residual disease in the bowel nor any evidence of metastases (ypT0N0M0).

3. Discussion

Recent studies have demonstrated remarkable response to ICB in patients with dMMR colorectal adenocarcinoma. In the NICHE study, patients with colon cancer were treated with a single dose of ipilimumab, a cytotoxic T-lymphocyte associated protein 4 (CTLA-4) inhibitor, and two doses of pembrolizumab, a programmed cell death protein 1 (PD-1) inhibitor, prior to conventional surgery [3]. All patients with dMMR cancers demonstrated a reduction in tumour volume following treatment and in 69 % of patients no evidence of residual disease was found in the surgical specimen [3]. In a separate study of patients with dMMR rectal cancers, complete clinical responses were seen in all 12 patients who were treated with dostarlimab, a PD-1 inhibitor, every 3 weeks for 6 months [4]. To date, none of these patients have demonstrated tumour regrowth and none have required any further treatment. The current case further highlights the efficacy of ICB in patients with dMMR colorectal cancers, whereby a patient with disseminated disease presumed to be incurable at diagnosis was effectively cured by this treatment alone.

However, this case also highlights some the challenges that the efficacy of ICB present in the management of these patients. Given the high rates of complete response to immune checkpoint blockade, one could argue that this treatment could form the basis of an organ-preserving strategy, whereby surgical resection is avoided altogether. Conventional surgery for colorectal cancer is associated with considerable risks, with peri-operative complications occurring in approximately 20 % and 0.6 % of patients dying within 30 days of surgery [6]. Surgical resection can also have significant impacts on sexual function and continence, which in turn can have significant impacts on quality of life [7], [8]. A major benefit of an organ-preserving approach is that the majority of these risks could be avoided whilst still effectively curing a patient of their disease. The case described above highlights these benefits, as one could argue that the patient was already cured before undergoing major surgery which ultimately led to their demise.

Despite these potential benefits, a major obstacle to the adoption of an organ-preserving approach is in accurately identifying patients who have achieved a complete response without the aid of pathological assessments. In the aforementioned study in patients with dMMR rectal cancers, clinical response was judged using both endoscopic and radiological assessments [4]. Whilst endoscopy appears to be a robust method for assessing the response of the primary tumour, radiological assessments of nodal status particularly in patients with colon cancers may be less reliable [9]. False positive findings (i.e. mistaken suspicion of residual disease) may have just as significant consequences as false negatives (i.e. mistaken suspicion of complete response), as demonstrated in the current case, whereby radiological suspicion of residual disease was a key factor in proceeding with cytoreductive surgery. Fluorodeoxyglucose positron emission tomography (FDG PET) may be of some benefit, with studies showing a close correlation between reduced metabolic activity and durable responses to immunotherapy [10]. However, false positives, whereby persistent metabolic activity did not correlate with residual disease have also been reported [11]. Monitoring circulating tumour DNA is an alternative method that is already showing promise in the surveillance of patients with colorectal cancer and may also have a future role in determining the response to immunotherapy [12].

There may also be concerns as to whether the neoadjuvant use of ICB has an impact on complications related to subsequent surgery. Wound complications have been described in patients with other cancers undergoing surgery after neoadjuvant ICB [13]. It is impossible to know if ICB played a role in the development of this patient's complications. However, the atypical presentation of their anastomotic leak, occurring 3 weeks after surgery, could raise suspicions that ICB may have had an influence. Of patients undergoing neoadjuvant ICB prior to surgery for colon cancer, 10 % developed an anastomotic leak, which is slightly higher than would be expected [3], [14]. However, the numbers in this study were small and whether ICB is genuinely associated with increased risks and/or atypical presentations of anastomotic leakage or other operative complications will have to be determined by future clinical trials.

The optimal duration of immunotherapy is a further unanswered question. Given that more than two-thirds of patients with dMMR colon cancer achieved complete responses after just two doses of ICB, one would expect an increased rate of complete responses with prolonged therapy. A longer course of PD-1 monotherapy was generally well tolerated in patients with dMMR rectal cancer, with no severe adverse events reported [4]. However, the concept of watchful waiting is less mature in the context of colon cancer than it is in rectal cancer, and it is unclear how acceptable this strategy will be for patients and clinicians alike. The need for serial endoscopic assessments is one potential obstacle, which in contrast to patients with rectal cancer, would require repeated cycles of bowel preparation. While many may consider that a small price to pay to avoid major surgery, some patients may prefer the finality of surgical resection. The importance of shared decision making with patients after careful discussion of different potential treatment strategies cannot be underestimated. It is also unclear whether ICB should be continued, and for how long, in patients who achieve a complete clinical response. Patients receiving ICB for various metastatic diseases have generally been treated for up to 2 years in the absence of progression or treatment limiting toxicities [15]. Shorter courses may not only limit toxicities but also reduce the financial implications of treatment although these potential benefits should only be investigated in the context of prospective clinical trials.

Finally, whether the long-term oncological outcomes of organ-preserving strategies are comparable to neoadjuvant treatment followed by conventional surgery remains to be seen. Local tumour regrowth (LTR) is well described in patients with rectal cancer treated with chemoradiotherapy alone, occurring in approximately 25 % of patients who achieve complete clinical responses [16], [17]. Whilst salvage surgery was feasible in the majority of patients who developed LTR, these regrowths were associated with increased risks of disease dissemination [18]. Inadequate responses to chemoradiotherapy appear to lead to expansion of resistant tumour cell populations, which may in part explain these increased risks for disease dissemination and provide a rationale for surgical resection to remove such resistant populations [19]. However, it may also be that inadequate responses reflect a more aggressive cancer biology, which would be associated with increased risks of relapse regardless of whether surgery was performed or not. Given that acquired resistance to ICB has been reported in other pathologies, one may expect regrowth to also occur to some extent following neoadjuvant ICB in patients with colorectal cancer and to present similar decision-making challenges [20].

4. Conclusion

ICB can lead to dramatic responses in patients with dMMR colorectal cancers. While these agents are likely to significantly alter the treatment paradigm of these patients in the future, their precise role remains to be defined. Major challenges remain in differentiating complete and partial responders and determining the indications for conventional surgery.

Consent

Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request.

Ethical approval

Not required.

Funding

None.

Author contribution

Concept and design - Henry G Smith, Anne Bodilsen, Lisbeth Rose, Rahim Altaf, Lene H Iversen, Line R Walker.

Data collection - Henry G Smith, Lisbeth Rose.

Manuscript preparation and review - Henry G Smith, Anne Bodilsen, Lisbeth Rose, Rahim Altaf, Lene H Iversen, Line R Walker.

Guarantor

Henry G Smith.

Research registration number

Not applicable.

Declaration of competing interest

None.