PD-L1 Scoring in NSCLC Varies Significantly

Although programmed cell death ligand-1 (PD-L1) is widely recognized as a key biomarker to predict the response of patients with non–small cell lung cancer (NSCLC) to immune-modulating drugs, a new report raises questions about significant variability in how pathologists score PD-L1 expression in patient samples.

The study is based on a survey of 32 pathologists in the United Kingdom. It found that despite using the same cutoff values, the proportion of patients who pathologists scored in the “negative,” “low,” and “high” categories varied significantly among centers. The authors say these findings point to an urgent need for more collaboration and consistency across the field. The study was published in Journal of Clinical Pathology.

The authors wrote that targeted therapies have helped to change the treatment landscape for people with NSCLC, including immune-modulating drugs targeting the protein cell death protein 1 (PD-1)/PD-L1 pathway.

The most common way to predict whether a patient will respond to immune-modulating drugs is to assess their PD-L1 expression levels, usually by calculating a tumor proportion score (TPS) based on the number of tumor cells with PD-L1 on their surface.

“Unfortunately, [PD-L1] is a fragile biomarker, compromised by its biological heterogeneity, variations in laboratory practice—including reluctance to use ‘cytology’ specimens for its assessment—and challenges in interpretation,” the investigators wrote.

They wanted to better understand how these challenges affect the experiences and performance of pathologists assessing PD-L1, so they created a questionnaire related to NSCLC PD-L1 testing practices and invited members of the Association of Pulmonary Practices to complete the questionnaire. Thirty-two of 44 invited pathologists responded to the survey. The results included both positive and negative findings.

“There was good consistency in practice and approach,” the authors found, “but there was also wide variability in the distribution of PD-L1 scoring.”

Forty-eight percent of centers reported their results using TPS, but 37% defined results as being within a “categorical range” based on the cutoff points of less than 1%, 1% to 49%, and 50% or higher.

Testing approaches were similar across centers. All of the centers tested second samples when the first specimen was inadequate, provided that such a sample was available. Second specimens were also frequently tested when the initial result was close to a cutoff point, the investigators reported.

“Occasionally, an oncologist would request testing of a further specimen from the same tumor site if the initial test on an adequate specimen had been ‘negative,’ but they were ‘running out of options’—the inference being that a second test might yield a higher score,” they wrote.

Overall, the proportions of samples scored as having negative, low, or high PD-L1 was broadly in line with previous studies (38%, 33% and 27%, respectively). Yet, the range of results within those categories alarmed the authors.

“Within each of the 3 categories defined by the usual ‘cutoff’ points, and which are often referred to as ‘negative’ (0% to < 1%), ‘low’ (1%-49%) and ‘high’ (≥ 50%), variation was wide at 23% to 70%, 10% to 60% and 15% to 36%, respectively,” the investigators wrote.

Previous trials have consistently suggested that the distribution of TPS scores ought to be relatively even across the 3 categories of negative, low, and high, they noted. And overall, “Broadly speaking, therefore, there is evidence from our survey that some centers may be ‘underreporting’ the PD-L1 TPS,” they wrote.

The authors said technological approaches, including the use of algorithms and machine learning, might be able to help solve the issue. However, they also said fixing the problem will require better collaboration.

“We suggest that a formal network is established of all laboratories engaged in PD-L1 testing of NSCLC with a view to sharing details of practice and data resulting from testing,” they said. “This would provide a basis for standardizing and improving practice and would carry an important educational component.”

Reference

Gosney JR, Peake MD, Kerr KM. Improving practice in PD-L1 testing of non-small cell lung cancer in the UK: current problems and potential solutions. J Clin Pathol. Published online January 5, 2023. doi:10.1136/jcp-2022-208643