Case presentation

CW is a Caucasian woman aged 70 years with history of asthma, sleep apnea (on continuous positive airway pressure), hypertension, and hyperlipidemia who presented to her primary care physician for her annual physical. She was otherwise well and specifically denied any gastrointestinal symptoms or unintentional weight loss. Per patient report and a review of outside records, she underwent a computed tomography (CT) scan of the chest without contrast for abnormal pulmonary physical examination findings. The CT scan revealed normal lung fields, but it was incidentally noted that she had thickening of the wall of the distal esophagus and the fundus of the stomach. Esophagogastroduodenoscopy was performed, which revealed a large, fungating, ulcerating mass within the gastric cardia extending proximally into the lower esophagus and distally into the gastric body (>3 cm in extension) along the lesser curvature. The pathology was consistent with a poorly differentiated adenocarcinoma. Biomarker status was not available, and the patient was referred to our institution for further management.

She underwent additional staging with positron emission tomography (PET) imaging, which revealed a metabolically active mass extending from the lower one third of the esophagus inferiorly to involve the gastroesophageal junction (GEJ) and proximal stomach, with a maximum standardized uptake value (SUV) of 25.9 (Figure 1A). Several other enlarged regional lymph nodes were noted without metabolic activity. No metabolically active distant metastatic disease was noted. She underwent staging laparoscopy, which revealed no evidence of peritoneal disease. The tumor seemed to involve the cardia and the proximal body of the stomach. Two small, superficial nodules were noted on the surface of the liver, one of which was resected at that time (and was identified as a cavernous hemangioma on histopathologic assessment). Cytology of peritoneal washings revealed benign mesothelial cells. Finally, a staging endoscopic ultrasound was performed, which revealed a hypoechoic, noncircumferential mass in the lower esophagus, cardia, fundus, and 5 cm into the body of the stomach along the lesser curvature. There was sonographic evidence suggesting invasion into the serosa. Six malignant-appearing lymph nodes were observed in the paracardial region (level 16), the gastrohepatic ligament (level 18), the celiac region (level 20), and the perigastric region (Figure 2A–F). This was staged as T3 (Tumor 3) with N2 lymph node status (uT3N2) by endosonographic criteria. During this time, the patient started to experience intermitted dysphagia to solids but was able to maintain adequate nutritional intake orally with supplementation. On histopathologic evaluation, the GEJ tumor biopsy revealed poorly differentiated adenocarcinoma (Figure 3). Biomarker testing specifically for mismatch repair (MMR) proteins was requested. Immunohistochemical staining for MMR proteins showed loss of expression of postmeiotic segregation increased 2 (PMS2) and mutL homolog 1 (MLH1) in tumor cells, but intact mutS homolog 6 (MSH6) and MSH2, consistent with MMR deficiency (dMMR). At baseline, a cell-free DNA liquid biopsy test (Guardant360^®; Guardant Health) was also performed. The highest variant allele frequency was 2.2%, with microsatellite instability high (MSI-H) detected (other genomic alterations noted on the cell-free DNA assay are listed in Table 1).

Medical oncologist’s perspective: Choice of systemic therapy and determining response

Choice of neoadjuvant chemoradiation versus perioperative chemotherapy—the ongoing conundrum in the treatment of GEJ cancers

There is no one gold standard for the treatment of locally advanced GEJ cancers, with both perioperative chemotherapy and chemoradiation (CRT) considered as appropriate treatment options. The CROSS trial (the Dutch Chemoradiotherapy for Esophageal Cancer Followed by Surgery Study) compared neoadjuvant CRT using five weekly doses of carboplatin/paclitaxel (CP) followed by surgery versus surgery alone. The addition of neoadjuvant CRT improved the negative (R0) resection rate (92% vs. 69%, respectively; p < .001) and improved median overall survival (OS) from 24 to 48.6 months (hazard ratio [HR], 0.68; 95% confidence interval [CI], 0.53–0.88; p = .003).¹ The FLOT4-AIO trial (ClinicalTrials.gov identifier NCT01216644) compared perioperative 5-fluorouracil, leucovorin, oxaliplatin, and docetaxel (FLOT) versus perioperative epirubicin-based triplet chemotherapy with fluoropyrimidine and platinum. Patients who received FLOT demonstrated an OS of 50 months compared with 35 months in the epirubicin-based chemotherapy arm (HR, 0.77; 95% CI, 0.63–0.94; p = .012).² These translated to very similar 5-year OS estimates for both neoadjuvant CRT (5-year OS, 47% in the CROSS study) and perioperative chemotherapy (5-year OS, 45% with FLOT), suggesting no clear winner. The recent Neo-AEGIS trial (ClinicalTrials.gov identifier NCT01726452) also drew similar conclusions (3-year OS, 55% with perioperative chemotherapy and 57% with neoadjuvant CRT).³

In this case, because there was a significant extension of the tumor in the stomach, we decided to start by using FLOT chemotherapy with perioperative intent (before biomarker study results were received).

Assessing treatment response with preoperative treatment

Where accurate objective measurements to assess treatment response are still needed, PET imaging has been explored as a marker for response to therapy. The CALGB 80803 trial by Cancer and Leukemia Group B (ClinicalTrials.gov identifier NCT01333033) was a randomized, open-label, phase 2 study of 257 patients with locally advanced esophageal and GEJ (Siewert types I and II) adenocarcinomas. The trial was intended to investigate whether PET imaging can be used to determine response to induction chemotherapy to then guide the chemotherapy backbone with subsequent CRT treatment. The two induction arms included either combined folinic acid (leucovorin), 5-fluorouracil, and oxaliplatin (FOLFOX) or CP and, in case of nonresponse on PET imaging (a <35% decrease in the maximum SUV), patients were switched to alternate chemotherapy with radiation. With induction FOLFOX, the pathologic complete response (pCR) rates were 40.3% and 18% in responders and nonresponders, respectively, which compared favorably with respective rates of 14.1% and 20% using induction CP. These pathologic outcomes translated to 5-year OS rates of 53% and 43.9% versus 37.5% and 40.4% for FOLFOX and CP responders versus nonresponders, respectively.⁴ Although the study showed that FOLFOX is a better choice for induction chemotherapy, it also demonstrated that, if the chemotherapy backbone was changed based on PET response, better survival rates can be achieved.

What we know about MSI-H gastroesophageal cancers

Molecular characterization of gastric cancers has revealed four different molecular subtypes: (1) Epstein–Barr virus-positive tumors, which demonstrate a higher prevalence of DNA hypermethylation and CDKN2A promoter hypermethylation, with 80% of them carrying a PIK3CA mutation; (2) microsatellite unstable (MSI-H) tumors, which have higher rates of hypermutation and MLH1 silencing as well as a predominance of genes in the immune signaling and mitotic pathways; (3) genomically stable tumors enriched in diffuse-type gastric cancers and predominated by mutations in CDH1 and RHOA; and (4) tumors with chromosomal instability, which are primarily of intestinal subtype and have frequent TP53 mutations as well as activation of RTK-RAS pathways. Based on this study, the prevalence of the MSI-H subgroup of gastric cancers is approximately 22%.⁵ In another molecular analyses of esophageal cancers, adenocarcinomas of the esophagus (EACs) had more molecular resemblance to gastric cancers than to squamous cell cancers of the esophagus. The majority of EACs showed significant overlap with the chromosomal instability subtype of gastric cancers. There were no EACs that showed MSI-H molecular features. However, GEJ cancers that were of not clear esophageal origin did have some that shared features with the MSI-H subtype of gastric cancers (approximately 8.3%).⁶ Contrary to these high prevalence rates of MSI-H in The Cancer Genome Atlas database, reports from real-world or retrospective studies suggest that MSI-H prevalence may be no more than 10% in gastric and GEJ cancers, whereas it is significantly lower in EACs (1%–5%).^{7, 8}

In the French single-arm, phase 2 NEONIPIGA study (ClinicalTrials.gov identifier NT04006262), 32 patients with locally advanced dMMR/MSI-H gastric/GEJ cancer were treated with neoadjuvant nivolumab (240 mg once every 2 weeks × 6 doses) and ipilimumab (1 mg/kg once every 6 weeks × 2 doses) followed by surgery and then adjuvant nivolumab (480 mg once every 4 weeks × 9 doses). Twenty-nine of these patients underwent surgery, and all had successful R0 resections. The pCR rate was 58.6%, and treatment-related grade 3 or 4 adverse events occurred in 19% of patients.⁹ In another single-arm, multicohort phase 2 study, the use of single high-dose tremelimumab 300 mg once and durvalumab 1500 mg every 4 weeks for three doses was investigated as definitive-intent treatment for dMMR/MSI-H, locally advanced gastric/GEJ cancers. Patients who did not have a complete response or had evidence of local recurrence were allowed to undergo surgery. Fifteen patients underwent surgery, and the pCR rate was 60%.¹⁰ These recent studies highlight the role of nonchemotherapeutic systemic therapy for patients who have dMMR/MSI-H, locally advanced gastroesophageal cancer.

Apart from these newer studies, when we look at some other select trials that did not exclude patients who had dMMR/MSI-H, we note that, for most of these cases in the locally advanced setting, surgery may be curative, and chemotherapy may not add a significant benefit. In the large phase 3 MAGIC trial (ClinicalTrials.gov identifier NCT01726452) of perioperative chemotherapy for gastric/GEJ cancers, of the 303 patients who had MSI results available, 20 (6.6%) their tumors were reported as MSI-H. The concordance of MMR and MSI results was approximately 97%. Patients who had dMMR showed a lack of pathologic response to chemotherapy. In the surgery-alone arm, patients who had MSI-H tumors had longer OS than patients who had microsatellite-stable or microsatellite stability-low tumors (OS not reached vs. 20.3 months; HR, 0.35; 95% CI, 0.11–1.11; p = .08). When treated with chemotherapy, OS was shorter in patients who had MSI-H tumor status compared with those who had microsatellite-stable or microsatellite stability-low tumor status (OS, 9.6 vs. 22.5 months; HR, 2.22; 95% CI, 1.02–4.85; p = .04).¹¹ In post-hoc analyses of the CLASSIC study (ClinicalTrials.gov identifier NCT00411229), which established capecitabine with oxaliplatin as an adjuvant regimen after resection of stage II/III gastric cancer, 6.8% of patients who were identified with MSI-H status did not derive any benefit in disease-free survival from chemotherapy (5-year disease-free survival, 83.9% with chemotherapy vs. 85.7% without chemotherapy; p = .931).¹² In the recent DANTE trial (ClinicalTrials.gov identifier NCT03421288), which compared FLOT plus atezolizumab versus FLOT, in a cohort of 23 patients (7.7% of the study population) with MSI-H status, the pCR rate was 63% with the combination versus 27% with chemotherapy alone.¹³

Our patient was treated before the results from the NEONIPIGA study (ClinicalTrials.gov identifier NCT04817826) or the INFINITY trial (ClinicalTrials.gov identifier NCT04817826) were presented. After three doses of FLOT, the PET response was modest (approximately32%); therefore, we decided to switch treatment to pembrolizumab 200 mg intravenously every 3 weeks. After four doses of pembrolizumab, the PET response was 81% from baseline and 73% since the start of immunotherapy treatment.

What is the best way to monitor these patients? Do all need adjuvant therapy?

There are two schools of thought when it comes to treating patients who have locally advanced, dMMR/MSI-H gastric/GEJ cancer. Although some believe that surgery alone may be sufficient; there are some that tend to err on the side of caution, especially when there is high nodal burden. In such cases, lessons learned from the NEONIPIGA and INFINITY trials, as well as some anecdotal cases like ours, shed light onto appropriate systemic therapy options. However, one begs to ask the questions, “Do we need adjuvant immunotherapy in all?”; and “Does the treatment have to be 1 year long?” It is hard to answer these questions today. However, recently, the growing use of circulating tumor DNA (ctDNA) for the purpose of minimal residual disease (MRD) testing has made us ask these important questions. Although complete reliability on these tests for clinical decision making is still ways ahead, some studies have truly shown value in this testing. In the IMvigor010 study (ClinicalTrials.gov identifier NCT02450331), which investigated adjuvant atezolizumab in patients who had resected urothelial cancers, those who were ctDNA-positive at the start of therapy had a poor prognosis, and there was a greater chance of ctDNA clearance in the atezolizumab arm (18%) versus the observation arm (4%; p = .0204).¹⁴ In a case series of three patients who had MSI-H colorectal cancer, after surgery, all of them were ctDNA-positive, which cleared with adjuvant immunotherapy, and they continued to demonstrate long-term disease-free status.¹⁵ In our case, we performed the first MRD test after surgery, which returned as negative. Given the pCR rate and negative MRD status, we elected not to offer any adjuvant immunotherapy to our patient CW. We have continued to monitor MRD every 3 months with her routine follow-up.

Radiation oncologist’s perspective: To treat or not to treat

We are beginning to build on the limited data guiding treatment decisions for locally advanced, dMMR gastroesophageal cancers. However, in extrapolating results from studies of rectal cancer, when neoadjuvant therapy is recommended, either upfront CRT or immunotherapy should be considered to avoid high rates of disease progression on traditional chemotherapy.¹⁶ Furthermore, some data suggest that patients with rectal cancer who have dMMR have a greater pathologic response to neoadjuvant CRT and may derive a benefit in recurrence-free survival; however, this has not been consistently reported.^{17, 18}

Regarding resectable gastric cancer, a retrospective review from two high-volume institutions comparing neoadjuvant chemotherapy strategies versus CRT demonstrated statistically significant improvements favoring CRT in completing prescribed therapy (91% vs. 63%; p < .001), an improved pCR rate (15% vs. 4%; p = .003), and was associated with improved OS (120 vs. 53 months; p = .015).¹⁹ Importantly, the MMR or MSI status was not reported, and the majority of patients did not receive FLOT chemotherapy, which is now considered standard of care.² Therefore, there is a lack of prospective, randomized data supporting the use of neoadjuvant CRT in gastric cancer versus neoadjuvant chemotherapy alone as we await the results of several current trials investigating this strategy, including TOPGEAR (ClinicalTrials.gov identifier NCT01924819) and CRITICS II (ClinicalTrials.gov identifier NCT02931890). It is important to consider that morbidity remains a significant impediment to this approach because the rate of grade 4 adverse events has approximated 21%.²⁰

In the future, investigations into proton therapy, such as on the NRG-GI006 trial for esophageal or gastroesophageal (Siewert I–II) cancer (ClinicalTrials.gov identifier NCT03801876) or magnetic resonance imaging-guided radiation, may help improve the therapeutic index of radiotherapy for this disease site.²¹ We soon may have more data to guide clinical decisions because there is currently an ongoing trial looking at a unique treatment regimen for selected subgroups patients with locally advanced gastric cancer, including those who have MSI-H status. The treatment involves two preoperative doses of pembrolizumab, followed by surgery, then adjuvant capecitabine/pembrolizumab with radiation followed by pembrolizumab alone to complete 1 year of treatment. In an interim analysis of 15 patients, six patients had MSI-H tumors; within this subgroup, there were two pCRs, and two thirds of the patients had downstaging.²² The final results from these trials will add insight into the role of radiation in gastric cancer, and especially in MSI-H gastric cancers.

Surgical oncologist’s perspective: Classification of tumor and surgical approach

Accurate classification of the tumor

The initial challenge in the management of patients with GEJ cancer is the correct classification of this disease. Anatomically, the tumors are divided based on the Siewert classification, named after the surgeon who led the group that originally proposed this classification.^{23, 24} This classification centers on the accurate localization of the epicenter of the mass—tumors with an epicenter 1–5 cm above the GEJ are referred to as Siewert I, tumors with an epicenter 1 cm above or 2 cm below the GEJ are considered Siewert II, and those with an epicenter 3 cm below the GEJ are referred to as Siewert III. Therefore, it is critical to achieve a best estimate of the true epicenter of the mass to assign a Siewert classification. To that end, we use a combination of imaging modalities and endoscopic procedures to estimate the probable origin and current location of the bulk of the tumor in relation to the anatomic GEJ, as opposed to simply focusing on the proximal and distal extent of the tumor and its midpoint.^{25, 26} This determination is not always straightforward and frequently is decided in a multidisciplinary approach in the context of a tumor board discussion. Once a Siewert classification is assigned, this informs potential treatment options as well as the nodal basins at risk for metastasis.^{24, 25}

Specifically, in our patient, the bulky nature of disease and the proximal and distal extent the mass resulted in defining it as a Siewert III tumor with extensive lower mediastinal and perigastric lymph nodes and with extension to the distal esophagus. However, with the initiation of immunotherapy and subsequent tumor regression, the epicenter of the mass proved to be centered at the true GEJ, which was more compatible with a Siewert II lesion. This highlights the need for continued multidisciplinary discussion and re-evaluation of these patients as they receive neoadjuvant therapy.

Surgical approach

The objective of the operation is to achieve a complete resection of the mass, obtaining negative margins proximally and distally while also removing the at-risk or involved lymph node basins in their entirety. After resection, one must be able to safely reconstruct the gastrointestinal tract. There is persistent debate regarding the optimal type of surgery for GEJ tumors; however, historically in the United States, we have treated Siewert I and II tumors as esophageal and Siewert III tumors as gastric.^{25, 27} Potential options for resection include total gastrectomy, proximal gastrectomy, and esophagectomy. Multiple studies have been conducted with the aim of identifying the optimal surgical resection type; however, ultimately, no particular approach has demonstrated superiority.^{25, 28, 29} Instead, surgical planning should be guided by aiming to achieve the aforementioned tenets of obtaining a margin-negative resection with safe reconstruction and complete oncologic lymphadenectomy. In addition, there appear to be clear benefits to performing resection using minimally invasive techniques—as opposed open surgery—when this is technically feasible. Specific benefits include reduced rates of postoperative complication and reduced length of stay. In particular, patients appear to benefit most with regard to a reduced rate of postoperative pulmonary complications.^30-32 However, it is clear that the use of minimally invasive and robotic techniques should be guided by experience and expertise because these procedures have a very steep learning curve.³³

Our case specifically highlights the need for multimodality tumor board review for initial assessment as well as continuous follow-up, because the tumor response ultimately informed the consensus discussion regarding optimal surgical approach. In our case, based on the tumor epicenter after initial therapy, we elected to perform a robotic-assisted Ivor–Lewis type esophagogastrectomy with resection of the proximal stomach. Lymphadenectomy included modified DII (levels 1, 2, 3, 4a, 7, 8, 9, 11p, and 20) and extensive lower mediastinal lymphadenectomy with thoracic levels 7, 8, 9, and 110 (para-aortic). In this way, we obtained both negative operative margins as well adequate lymphadenectomy, feeling more confident in the ypT0N0 outcome (complete pathologic tumor and lymph node regression after initial therapy), and removal of at-risk nodes.

Discussion

The DNA MMR is a highly conserved system that identifies and repairs mismatched nucleotides occurring during genetic recombination or acquired because of physical or chemical damage. MMR ensures stability and integrity of the genome and prevents insertions or deletions of abnormal genomic material. The four key genes comprising the MMR system include: MLH1, PMS2, MSH2, and MSH6. DNA mismatched base errors can occur because of single-base mismatch or erroneous insertion or deletion changes. The heterodimer of MLH2/MSH6 binds to these DNA mismatched base errors, whereas the heterodimer of MLH1/PMS2 excises and resynthesizes corrected base errors at these sites. Protein expression of MLH1, PMS2, MSH2, and MSH6 can be assessed using immunohistochemistry (IHC), and loss of expression or dysfunction of one or more of these is called dMMR. Abnormal function of MLH1 or MSH2 does not form the heterodimers MLH1/PMS2 or MSH2/MSH6, and this results in the degradation of PMS2 or MSH6.³⁴ Therefore, IHC will show loss of staining for both proteins. Lynch syndrome is an autosomal-dominant hereditary condition caused by germline mutations in MMR genes or by a heterozygous germline deletion in the epithelial cell adhesion molecule (EPCAM) gene that regulates MSH2 gene expression.^{35, 36} The genomic alterations that cause Lynch syndrome include loss of MSH2, significant mutations in the MLH1 or MSH2 genes, MLH1-methylation inactivation, and transcriptional silencing.³⁷ Not all four genes are affected similarly in Lynch syndrome: MLH1 accounts for 42%–50%, MSH2 accounts for 33%–39%, MSH6 accounts for 7%–18%, and PMS2 accounts for <7% of deletion mutations.³⁸

Either germline mutations in the MMR genes or spontaneous hypermutation alterations can induce MSI, which comprise more than 100,000 areas of short tandem repetitive DNA sequences. Standard MSI testing includes two mononucleotide repeats (BAT25 and BAT26) and three dinucleotide repeats (D5S346, D2S123, and D17S250). MSI-high is defined when two or more repeats are altered; and, if only one repeat is altered, it is termed MSI-low. There is 90%–95% concordance between dMMR and MSI-H.³⁹ dMMR or MSI-H is seen most commonly in endometrial cancer of the uterine corpus (17%–31%), colon adenocarcinoma (6%–19%), gastric adenocarcinoma (9%–19%), rectal adenocarcinoma (approximately 5%), adrenocortical carcinoma (approximately 4%), and uterine carcinosarcoma (3.0%–3.5%). The frequency of dMMR is higher in early stage than in late-stage cancers.⁴⁰ MSI-H has been detected in 10%–20% of gastric cancers, and these patients have improved survival compared with patients who have gastric cancers that are microsatellite-stable.^{41, 42}

Follow-up

Based on multidisciplinary assessment, the initial recommendation was to start CW on perioperative systemic therapy with FLOT and growth factor support. The first treatment was started at 4 weeks from initial diagnosis. Because biomarker testing was still pending at the time, a liquid biopsy was obtained at the start of treatment. Results of the liquid biopsy test were obtained after the first FLOT treatment, and the key finding was detection of MSI-H. In the meantime, biomarker testing results were also obtained. The IHC for MMR proteins revealed loss of expression of PMS2 and MLH1 in tumor cells and intact MSH6 and MSH2 (Figure 3). CW’s case was presented at the multidisciplinary tumor board. Because of the finding of dMMR, consideration for upfront surgical resection was discussed. However, because of nodal disease and the evolving role of checkpoint inhibitors to treat MSI-H solid tumors, the tumor board recommended continuing FLOT for total three doses and repeat imaging to assess for response. A repeat PET was obtained, and this showed persistent metabolic activity at the primary site with an SUV response of approximately 32% (Figure 1B). No distant sites of metastatic disease were noted.

CW was finding it harder to tolerate FLOT; and, based on a very modest response on PET imaging, the decision was made to switch to the checkpoint inhibitor pembrolizumab 200 mg intravenously every 3 weeks for four treatments. Repeat PET imaging in November 2020 showed a small amount of moderate uptake in the distal esophagus with an SUV of 4.8 (Figure 1C). After three additional doses of pembrolizumab, a repeat PET image in January 2021 revealed that the SUV was stable at 4.7. In February 2021, the patient underwent a robotic-assisted laparoscopic Ivor–Lewis esophagectomy with modified D2 lymphadenectomy and feeding jejunostomy placement. An endoscopic evaluation at the time of resection revealed an ulcerated mass starting at 2 cm proximal to the GEJ and extending 2 cm into the cardia of the stomach. The abdominal lymphadenectomy encompassed the hepatic, periportal, celiac, splenic take-off, and left gastric lymph nodes and the lesser curvature and cardia lymph nodes. There was a resection of the fundus and a significant portion of the stomach, with creation of a short, 4-cm diameter gastric conduit. An intraoperative frozen section was obtained of the gastric margin before proceeding to the thoracic portion of the procedure. The lower mediastinal, right inferior pulmonary, periaortic, level 7, and right level 8 were resected with a right thoracic anastomosis. She tolerated the surgery well, and she was discharged home on postoperative day 7 after a negative upper gastrointestinal test with double contrast fluoroscopy. At the time of discharge, her nutrition was an oral liquid diet and supplemental enteral feedings. Pathologic evaluation revealed a pCR. The patient is currently on surveillance and continues to do well. She has had expected weight loss after surgery but has no long-term side effects. She is currently being monitored according to standard guidelines in addition to serially performed MRD testing.

Based on dMMR and MSI-H detection on somatic liquid biopsy, CW was referred for germline testing. Both sides of her family were of non-Ashkenazi Jewish descent. Her family history included her father diagnosed with an unknown cancer at the age of 49 years (history of tobacco and alcohol use) and a half-sister who was diagnosed at the age of 64 years with lung cancer (possible asbestos exposure). Therefore, there were no patterns of hereditary genetic syndrome based on her family history. The BRCA2 T3085fs mutation noted on somatic liquid biopsy is a pathogenic mutation when seen on germline testing. Therefore, the patient decided to proceed with the 84-gene Multi-Cancers panel (Invitae Corporation) along with all preliminary evidence genes (156 in total) offered by Invitae. This testing did not detect any known deleterious mutations. The full results were: ALK c.2012C>T (p.Pro671Leu) heterozygous variant of uncertain significance, ATM c.1166 T>C (p.Ile389Thr) heterozygous variant of uncertain significance, and FANCE c.401 G>A (p.Arg134His) heterozygous variant of uncertain significance. Neither of the BRCA1 mutations identified on somatic testing was detected, suggesting that these were only somatic in origin. Therefore, we concluded that the patient has a sporadic case of dMMR or MSI-H gastroesophageal cancer.

Conclusions

With this case, we wanted to highlight the nontraditional, yet multidisciplinary, treatment of locally advanced, dMMR/MSI-H gastric cancer. These patients comprise a distinct subgroup and, although their cancers may not be very common, they do deserve to be recognized. At very early stages when there is no nodal disease or the nodal burden is very low, an upfront surgical approach can be reasonable. However, in patients with more locally advanced disease, it is prudent to consider neoadjuvant treatment. With a growing body of literature in both locally advanced and metastatic settings for dMMR/MSI-H gastric/GEJ cancers using immunotherapy, chemotherapy-free regimens with checkpoint inhibitors, or at least chemoimmunotherapy, certainly must be considered. If neoadjuvant immunotherapy is offered, do all patients need adjuvant therapy? We believe that integrating the use of ctDNA in this setting will help guide the use and duration of adjuvant treatment. The role of radiation therapy in these patients is yet to be determined. Although we most commonly hear that dMMR/MSI-H cancers are associated with Lynch syndrome, it is important to refer these patients for germline testing. Like our patient, sporadic cases may be seen. Regardless, our patient had a good response to immunotherapy. Therefore, it is too soon to draw any conclusions regarding whether sporadic cases versus those associated with Lynch syndrome would have differing sensitivities to immunotherapy. We strongly encourage MMR and/or MSI testing for all patients who are newly diagnosed with locally advanced gastric or GEJ cancer.