Atypical pattern of response in rectal cancer after neoadjuvant pembrolizumab treatment: a case report, literature review, and proposed management model

Abstract

Background: Immunotherapy is the first-line treatment in patients with advanced microsatellite instability-high (MSI-H) or deficient mismatch repair (dMMR) colorectal cancer (CRC). Although immune checkpoint inhibitors (ICIs) for locally advanced rectal cancer (LARC) are not yet a standard, the results are very encouraging and raise the question of whether patients with clinical complete response (cCR) could receive nonoperative management (NOM). However, different patterns of response have challenged management strategies. Case Description: A 34-year-old woman diagnosed with dMMR LARC started treatment with capecitabine 2,000 mg/m2 on day 1 to 14 and oxaliplatin 130 mg/m2 on day 1 and every 21 days. Magnetic resonance imaging (MRI), performed three cycles later, showed local progression of the primary rectal lesion, which at that time had new peritoneal reflex involvement. A new hepatic lesion in segment V was observed. Due to disease progression, she was administered pembrolizumab 200 mg every 21 days. After three cycles, a discordant radiological response was observed on a new MRI scan that showed a complete response of the liver lesion and magnetic resonance tumor regression grade (mrTRG) 1 in the rectum. However, new involvement of the mesentery and enlargement of the regional lymph nodes (LNs) were also evident. A new colonoscopic biopsy was performed, showing no cancerous cells. She underwent surgery on the rectum and liver lesion. Pathology showed a complete response of the rectal wall and liver lesion, but 1 of 22 LNs was positive for adenocarcinoma (ypT0 N1 M0). The patient continued on pembrolizumab, and 14 months after surgery, she had not relapsed. Conclusions: Neoadjuvant immunotherapy for rectal cancer requires new recommendations for the assessment of clinical response. Pseudoprogression should be ruled out as an atypical response before deciding on surgical treatment. We propose an algorithm to address pseudoprogression in this setting.