PD-L1 Testing

PD-L1 testing measures the amount of programmed death ligand 1 on tumour cells and immune cells in a tumour section, by immunohistochemistry. PD-L1 binds PD-1 on T cells and switches them off, so a tumour that expresses it can hide from the immune system. Drugs that block PD-1 or PD-L1 release this brake, and the test helps to choose patients who are more likely to respond.

The test is done on formalin fixed, paraffin embedded tumour tissue, with an antibody and a staining platform that have been validated for the drug and the tumour type. Results are reported as a score, and the score cut offs differ between tumour types and drugs. Turnaround is 1 to 3 days.

Two scores are used most often. The tumour proportion score (TPS) is the percentage of viable tumour cells with partial or complete membrane staining. The combined positive score (CPS) is the number of PD-L1 staining cells (tumour cells, lymphocytes and macrophages) divided by the total number of viable tumour cells, multiplied by 100 and capped at 100.

PD-L1 expression is found in about 25 to 30% of non-small cell lung cancers at a TPS of 50% or more, and in a large proportion of gastric, head and neck, cervical, urothelial and triple negative breast cancers at lower cut offs. It is an imperfect predictor, since some patients with no expression respond and many with high expression do not.

Indications.

• Advanced non-small cell lung cancer without an EGFR mutation or ALK rearrangement, to select first line immunotherapy alone or with chemotherapy.
• Gastric, gastro oesophageal junction and oesophageal cancer (CPS thresholds).
• Head and neck squamous cell carcinoma (CPS thresholds).
• Triple negative breast cancer (CPS of 10 or more).
• Cervical, urothelial, vulval and some other cancers, depending on the drug.
• Re-testing when a new line of immunotherapy is considered and no recent result exists.

Not indicated in cancers where immunotherapy is given irrespective of expression, or where a drug has no PD-L1 requirement. It is also not a substitute for testing of mismatch repair status and driver mutations.

Specimen requirements.

• Formalin fixed, paraffin embedded tumour (biopsy or resection), fixed for 6 to 72 hours in 10% neutral buffered formalin.
• At least 100 viable tumour cells for scoring. Cytology cell blocks are acceptable where they have been validated.
• Fresh cut sections. PD-L1 staining fades in sections stored for months, and slides are cut shortly before staining.
• Bone and decalcified tissue are poor material, and acid decalcification is avoided.
• Clinical details: tumour type, stage, previous treatment, and the drug being considered.

Normal appearances.

Normal tissue and tonsil tissue used as a control show weak or moderate staining in germinal centre macrophages and some epithelium. Tumours with no expression show none on tumour cells, and internal controls (macrophages, lymphocytes) confirm that the stain worked.

Interpretation.

• Lung (TPS): under 1% is negative, 1 to 49% is low expression, and 50% or more is high expression. High expression is linked with the largest benefit from immunotherapy alone, and tumours with 1 to 49% are usually treated with immunotherapy combined with chemotherapy.
• Gastric and oesophageal (CPS): thresholds of 1, 5 or 10 are used, depending on the drug and the setting.
• Head and neck (CPS): thresholds of 1 and 20.
• Triple negative breast cancer (CPS): 10 or more.
• Cervical and urothelial cancer: CPS thresholds of 1 or more, in settings set by the drug approval.
• Staining on tumour cells only, immune cells only, or both, is reported in the way that the scoring system defines.
• A result that is negative or low does not exclude benefit, and it does not remove immunotherapy as an option when the drug has no required threshold.
• Heterogeneous staining: the score reflects the whole section, and a hotspot is not scored alone.

Critical values and findings needing immediate action.

• High expression (TPS of 50% or more) in a patient with advanced lung cancer and no EGFR, ALK or ROS1 alteration: urgent referral to the oncology team for first line treatment.
• A driver alteration found in a PD-L1 positive tumour: targeted treatment is usually preferred, since immunotherapy has less effect and can raise toxicity when a targeted drug is given soon after.
• Failed or inadequate sample: repeat testing on another block or a new biopsy without delaying the choice of treatment.

Limitations and artefacts.

• Different antibody clones, staining platforms and scoring systems are approved for different drugs, and results are not freely interchangeable.
• Pre-analytical factors (delay to fixation, under fixation, over fixation, decalcification, old sections) give false negative results.
• Heterogeneity: expression varies within a tumour and between primary and metastatic sites, and a small biopsy may under represent it.
• Scoring is subjective, with variation between observers close to a cut off, and training and external quality assurance are needed.
• Staining of necrotic tissue, pigment and macrophages in tumour nests is misread as tumour staining.
• Expression changes with treatment, and a test done years earlier or before chemotherapy may not represent the current tumour.
• It is a partial biomarker, and tumour mutation burden, mismatch repair deficiency and immune cell infiltrates add information.
• Turnaround, cost and access limit use.

Special groups.

• Patients with EGFR-mutated or ALK-rearranged lung cancer: PD-L1 expression may be high, and immunotherapy gives little benefit.
• Patients with small biopsies: material is saved for molecular testing, and a stepwise plan is needed.
• Patients with autoimmune disease or organ transplants: immunotherapy has particular risks, and the decision is individual.
• Patients with prior immunotherapy: re-testing has uncertain value.
• Regions: PD-L1 testing is available in few laboratories, and the cost of immune checkpoint drugs often limits access. Where treatment is not available, testing has little value, and resources go first to hormone receptor, HER2 and mutation tests that direct cheaper treatment.

Follow up tests.

• EGFR, ALK, ROS1, BRAF, KRAS, MET, RET and NTRK testing in lung cancer.
• Mismatch repair immunohistochemistry or microsatellite instability testing.
• Tumour mutation burden on a sequencing panel.
• HER2 testing in gastric and breast cancer.
• Repeat biopsy at progression.
• Baseline thyroid, adrenal, liver and renal function before immunotherapy.
• Staging imaging.
• Multidisciplinary meeting for treatment planning.