I am a biologist and I conducted a study on cell lines, studying markers of resistance for various anticancer molecules, cytotoxic and targeted. I get the gene SLFN11 as a clear biomarker of sensitivity/resistance across several cell types and molecules. I noticed that the literature is full of evidence that the expression of this gene is associated with survival and response. E.g.,

So, why is it not used in clinical practice? It seems to be relevant for many cancer types, including NSCLC, breast, ovarian.

  • 2
    Hi @SebDL and welcome to Medical Sciences. Can you provide links to some of that literature about SLFN11 being a biomarker correlating with cancer survival? What cancers is it relevant to and what treatments does it predict response for?
    – Chris
    Dec 5, 2023 at 12:32
  • 1
    For example, I found this that discusses SLFN11 expression in relation to small cell lung cancer. This article is from 2021 and states more work is needed. Also there is often a lag of a few years until something is adopted clinically.
    – Chris
    Dec 5, 2023 at 12:37
  • 5
    As a clinical test, its sensitivity and specificity, and the interpretation of test results matter a lot, not to mention the administrative cost etc. All these are beyond purely biological findings that it is a biomarker.
    – wilsonw
    Dec 5, 2023 at 12:58
  • 1
    It looks like the oldest paper referred to in the first reference is 2016. One of the papers refers to a study of 7 people. I think a reasonable expected preclinical -> clinical pipeline is on the order of 20 years. The mRNA COVID vaccines are quite an exception to the ordinary pace of things. You may also need to ask the question: "If we had access to this biomarker data for a single subject, how would it change treatment strategies?" Has anyone done a large trial showing that use of this biomarker to change a treatment decision actually (rather than theoretically) improves outcomes?
    – Bryan Krause
    Dec 5, 2023 at 15:30
  • 4
    Did you read the article in your second link? It specifically states Despite the abundance of preclinical evidence that SLFN11 expression is predictive of many chemotherapy agents and PARPi targeting DDR, no prospective trials using SLFN11 as a predictive biomarker have been completed. To our knowledge, the randomized Phase II trial of atezolizumab plus talazoparib versus atezolizumab alone in the maintenance setting for SLFN11-positive extensive-stage SCLC (NCT04334941) is the first clinical trial selecting patients with SLFN11-positive tumors. ...
    – bob1
    Dec 5, 2023 at 20:15

1 Answer 1


Here are some general ideas regarding factors affecting implementation of biomarkers into a clinical setting.

  • Biomarker performance
    • biomarkers are often discovered with specific experimental designs, where the prevalence of disease is completely different from the actual prevalence, leading to overestimations of e.g. the positive predictive value or accuracy in the actual population
    • a biomarker identified in a single study with very sensitive techniques (e.g., mass spectrometry) may not be robust or replicable
    • performance must be evaluated in comparison to other biomarkers (see costs)
  • Technical implementation
    • the identification of biomarkers is often made with techniques that are not usable as is in a clinical setting (full sequencing, mass spectrometry, ...), therefore they are often brought to other types of platforms, which is a technical challenge but also may not support the initial findings
    • is associated with additional costs
    • the test matrix (type of sample in which the biomarker can be identified) must be available (i.e., if tumor excision is required to run the test immunohistochemistry or other tests with validated prognostic value will likely be performed anyway)
  • Cost/risk/benefit
    • must consider plausible scenarios of sampling, analysis and subsequent consequences
    • must provide an advantage in the above compared to existing biomarkers
  • Actual need
    • biomarker in comparison to other established predictors
    • relationship between the biomarker and the treatment approach/outcome
    • proportion of cases which would benefit from this additional biomarker
  • Legal considerations
    • patents (existing ones or lengthy procedure to obtain patent rights)

In the case of SLFN11 I would think that the implementation is not very difficult but that the benefits need to be demonstrated for each cancer type on a case-by-case basis and in comparison to the current standards for prognostic grading of these types of neoplasias. The review publications you linked suggest that cancers of a single type may express SLFN11 to different degrees, which will predict the efficacy of DNA-damaging agents in the treatment of those cancers.

Here are a few points one would need to evaluate for each cancer type:

  • Are DNA-damaging agents typically used (or sufficiently promising) to treat such cancers?
  • If yes,
    • does the rate of non-responsive neoplasias warrant further investigations into SLFN11 or is it very low?
    • are there alternatives to DDAs for treatment and would a SLFN11 test justify one of those alternatives with respect to benefit/risks?
    • will follow-up be performed anyway and is it performant at detecting non-responders?
    • where (chronologically) should SLFN11 testing be implemented into the standard procedure, since SLFN11 expression may vary with time and depending on previous treatments?

So the answer to your question likely requires a lot of experts of different fields to work together and that takes a lot of time...

  • Thanks for the detailed answer!
    – SebDL
    Dec 7, 2023 at 9:23
  • Hi @SebDL, you're welcome. While the definitive answer to your question is cancer specific and I don't have sufficient knowledge to answer more specifically, if the answer is satisfying, I would appreciate it, if you could accept it by ticking the check mark on the left :)
    – jkd
    Dec 7, 2023 at 11:00
  • Upvoted, but just wanted to express, really nice answer! Also, too soon to tell, but perfectly good answers often don't get accepted-ticked on SE sites, sometimes simply because it's not what the OP wants to hear. Dec 7, 2023 at 15:20
  • Thanks @anongoodnurse, I appreciate the feedback :)
    – jkd
    Dec 8, 2023 at 9:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.