Biomarkers in Gastrointestinal Cancers

PD-L1 expression, HER2 status, and claudin18.2 status are already being taken into account in the choice of treatment of patients with colon or gastric cancers, while new markers such as FGFR2b are on their way into clinical practice. At an OncoViews webinar, PD Dr. Sara De Dosso, Istituto Oncologico della Svizzera Italiana, Bellinzona, PD Dr. Heather Dawson, Institute for Tissue Medicine and Pathology, University of Bern, and Prof. Dr. Thomas Greuter, Ospedale Regionale di Lugano, Civico, discussed the topic of gastrointestinal biomarkers from the perspective of three disciplines. The following text is an abridged version.

Optimization of sample collection

De Dosso: Several targeted substances have become available for the treatment of gastrointestinal cancers, and many more will be available soon, which is why molecular testing is of increasing importance. The first question focuses on how to ensure high quality of the biopsy specimens. Are specific protocols being used during endoscopies to improve tissue yield?

Greuter: We try to take as many samples as possible to maximize tissue yield, and to biopsy target lesions. For lymph nodes, we no longer perform fine needle aspirations, but rather fine needle biopsies with larger needles. Perception can be deceiving here; even if we believe that the amount is sufficient, the pathology report sometimes states that it is not. Endoscopic tools such as narrowband imaging help to delineate the cancerous part of a lesion. However, standardized protocols for biopsies are completely lacking.

De Dosso: For us as oncologists, it is essential that gastroenterologists are aware that endoscopy is not just about making a diagnosis.

Greuter: I think that gastroenterologists lack awareness in this respect. Oncology is developing much faster than gastroenterology does, which is why keeping track on markers and novel treatments is challenging for gastroenterologists. Continuing education plays a key role here.

Preanalytics and reflex testing

De Dosso: How are samples being handled by pathologists to ensure certain standards?

Dawson: Basically, we have to work with the tissue we receive. There are highly standardized protocols for biomarker testing. Antigenicity must be preserved, as all these biomarkers are immunohistochemical staining markers. The tissue must be fixed correctly, avoiding under-fixation (< 8 hours) and over-fixation (> 72 hours). It is also necessary to put the sample into formalin as quickly as possible after removal, as degradation begins as soon as the tissue is removed from the body. A key aspect is therefore the correct handling of pre-analytical phase. There is sufficient evidence and guidelines for this. The fact that the initial focus is solely on the diagnosis leads to a lack of awareness not only in the gastroenterology community, but also in the pathologist community. Routine pathology reporting does not indicate whether a sample is adequate for biomarker testing, which is why gastroenterologists receive little feedback in this regard.

“It is very important to be aware of the molecular status at the start of first-line treatment”

PD Dr. Sara De Dosso, Istituto Oncologico della Svizzera Italiana, Bellinzona

De Dosso: At our institute, reflex testing of the microsatellite status is primarily based on the mismatch repair protein. In case of doubt, next-generation sequencing (NGS) is used. Furthermore, the PD-L1 and EBV status are determined as soon as a diagnosis of adenocarcinoma of the stomach has been established. How is this being handled at other clinics in Switzerland?

Dawson: This is topic of the debate. Professional societies increasingly recommend that all biomarkers should be collected upfront. On the other hand, the list of biomarkers is growing for various types of carcinomas, which increases the workload for pathologists and laboratories. In Switzerland, we can handle this, but in other countries it is certainly a big issue.

De Dosso: The discussion now revolves around whether HER2 testing should be performed as a reflex test. In my view, it is very important to be aware of the molecular status at the start of first-line treatment for gastric cancer so that the best strategy can be discussed immediately after a patient has been diagnosed with metastatic disease.

Greuter: Sometimes that is not feasible. For example, a tumor may be detected in a patient with gastric bleeding, but a biopsy cannot be performed because the patient is on anticoagulation.

How reliable is MMR testing?

De Dosso: One alarming finding was the high rate of misdiagnosis regarding microsatellite status in the two phase III KEYNOTE-177 and CheckMate 8HW studies.^1,2 In 15% of colon cancer patients, the central review did not confirm the results of local testing.³

Dawson: I also saw this data and I was shocked. This figure does not match the information found in the literature. In the studies, the concordance rate between PCR and immunohistochemistry for the microsatellite status or mismatch repair (MMR) proficiency is much, much higher. However, this analysis did not provide much detail on the type of testing, which was also criticized in a letter to the editor. A reply was then published, essentially stating that we currently lack the depth of data to explain this phenomenon. One reason for this may be a high incidence of tissue artifacts and necrosis in the biopsy material, or the fact that some institutes use only two immunohistochemical markers as a screening tool, instead of four. Finally, many preanalytical factors can influence the result. The threshold is also an issue: at what point can the tumor be safely classified as MSI high or MMR deficient according to immunohistochemistry? And what is the threshold at which to perform PCR?

Accuracy of PD-L1 assessment

De Dosso: Are there still institutions that report the PD-L1 status without the CPS score that is important for oncologists?

Dawson: At least in Switzerland, I am not aware of any. This should no longer be the case. In Switzerland, there is actually a lack of guidelines regarding the development and validation of PD-L1 tests. I am not sure if every laboratory is aware of how this should be done. Other scientific societies have guidelines for establishing, validating, and reporting of biomarkers in gastrointestinal cancer. With regard to PD-L1 reporting, for example, a recent publication has shown a very high degree of variability between laboratories.⁴ There are many knobs that can be turned and that have a significant impact on the result.

“In Switzerland, there is actually a lack of guidelines regarding the development and validation of PD-L1 tests. I am not sure if every laboratory is aware of how this should be done.”

PD Dr. Heather Dawson, Institut für Gewebemedizin und Pathologie, Universität Bern

De Dosso: The degree of PD-L1 expression is crucial for the choice of treatment. Given the availability of zolbetuximab, PD-L1 expression is the deciding factor in patients with gastric cancer who are both claudin18.2- and PD-L1-positive, as immunotherapy is less effective at lower CPS scores. These scores must of course be reliable.

Dawson: Many factors are relevant for the accuracy of CPS classification. High intratumoral heterogeneity plays a role, as do preanalytical and analytical variability. Interpreting PD-L1 requires counting a lot of cells and the ability to accurately distinguish cell types not all of which are visible at once. This is difficult and time-consuming. Moreover, the AI tools we currently have are not ideal, but we will certainly be able to overcome these problems within the next five years and to achieve generalizability of results across laboratories.

De Dosso: Should the regulatory perspective also play a role in PD-L1 testing, given that checkpoint inhibitors have been approved for certain CPS scores?

Dawson: This is difficult for pathologists because PD-L1 has become relevant for many carcinomas and there is now a jungle of scores, regulatory definitions, cutoffs, etc. A highly commendable German website, which requires users to log in, lists all regulations and scores relevant for certain drugs according to organ systems.⁵

Greuter: Clinical information can lead to a diagnostic bias.

Dawson: This is true, but I still believe it is very important to know what the consequences of a certain diagnosis would be.

Targeted clinical development of biomarkers

De Dosso: With respect to the role of the pharmaceutical industry in the development of biomarkers, I would prefer academic development, with samples being provided based on large clinical trials.

Dawson: In Switzerland, academic centers should show more interest in contributing to clinical trials. Centralized biomarker development is fine, but we should perhaps be more actively involved in the development of biomarker assessment and scoring. There should also be more openness toward specific NGS platforms if it can be demonstrated that the results are valid and of high quality. This could also lead to greater involvement of academic centers.

De Dosso: Unfortunately, large phase III studies often have methodological weaknesses with regard to biomarkers, such as the lack of pre-planned analyses. Drugs may therefore be approved on the basis of post-hoc and subgroup analyses. When designing a study with a targeted drug or immunotherapy, one should already have an idea of the best target group based on preclinical or early clinical studies.

Dawson: Other tissue-based markers that have already been established could also be included in order to define subgroups beyond clinical characteristics.

Improving the exchange of information

De Dosso: New biomarkers and new drugs are on the horizon; will we need even more tumor tissue for testing?

Greuter: The amount of material will forever be an issue. We will therefore need standardized protocols in the future with regard to the type and quantity of tissue. Moreover, we need continuous feedback from the pathologists as to whether the material is sufficient. Realistically speaking, we are moving toward a second endoscopy to obtain the desired amount of material after the diagnosis has been made during the first endoscopy and the patient has been discussed in the tumor board. Hopefully, this will not give rise to issues regarding reimbursement.

“The amount of material will forever be an issue.”

Prof. Dr. Thomas Greuter, Ospedale Regionale di Lugano, Civico, and Università della Svizzera italiana.

De Dosso: What will be the biggest challenges in the future in the context of new biomarkers, and how can we support pathologists?

Dawson: Every new biomarker requires appropriate training to ensure high-quality testing at all levels.

De Dosso: A final point of discussion is the need for interdisciplinary exchange between medical oncologists, surgeons, radiation oncologists, gastroenterologists, and pathologists. There is often little communication beyond multidisciplinary team meetings. What channels could be used to improve the sharing of news?

Greuter: It is very important to constantly receive feedback. I know institutions where gastroenterologists and pathologists evaluate tissue sections together for educational purposes, which benefits both doctors and patients. But the implementation of such formats, which come at the expense of direct patient care, must be a top-down decision.

Dawson: The exchange can take place at all events that bring us together anyway. However, everyone is pressed for time and needs to earn continuing education credits. It therefore makes sense to focus on specific interdisciplinary sessions. And I would like to encourage the readers to communicate with their colleagues so that we can learn from each other.

Watch the full video to explore more on GI biomarkers with Drs. Sara De Dosso, Heather Dawson and Thomas Greuter. 

References

1. André T, Shiu KK, Kim TW, et al. Pembrolizumab in microsatellite-instability–high advanced colorectal cancer. N Engl J Med. 2020;383(23):2207-18.
2. André T, Elez E, Van Cutsem E, et al. Nivolumab plus ipilimumab vs chemotherapy as first-line treatment for microsatellite instability-high/mismatch repair-deficient metastatic colorectal cancer: First results of the CheckMate 8HW study. J Clin Oncol. 2024;42(3_suppl):LBA768.
3. Marinelli D, Sabatini A, Bengala E, Ciurluini F, Picone V, Santini D, et al. Systemic treatment of mismatch repair deficient/microsatellite instability-high metastatic colorectal cancer—single versus double checkpoint inhibition. ESMO Open. 2024;9(6):103483.
4. Robert ME, Mullaney J, Roden AC, et al. High Interobserver Variability Among Pathologists Using Combined Positive Score to Evaluate PD-L1 Expression in Gastric, Gastroesophageal Junction, and Esophageal Adenocarcinoma. Mod Pathol. 2023;36(5):100154.
5. Qualitätssicherungs-Initiative Pathologie QuIP GmbH, quip.eu (last accessed: 25th November 2025).