You are here

How I select the right treatment for the right patient: histological and molecular profiling challenges


2 Fernando Lopez-Rios

Fernando López-Ríos

Dr López-Ríos reviews how to select patients for treatments based on histology and molecular profiling. He highlights the current problems with pre-analytics and histology including small sample sizes (of DNA, liquid biopsies, RNA) and sampling late in diagnosis. To improve this, he suggests bringing patients directly to the laboratories for sample collection, involving a multidisciplinary team (particularly in the case of rebiopsies), and selecting the best sample from each patient on which to perform the assays. Dr López-Ríos notes that the risk of false negatives and false positives are increased with too few tumour cells, and highlights the difficulty in determining the differences between adenocarcinomas and squamous cell carcinomas.

Dr López-Ríos then reviews sections of the proposed CAP/IASLC/AMP Molecular Testing Guideline for Selection of Lung Cancer Patients (2016 draft recommendations).

When considering the analytical phase of profiling, there are too many choices available that can ultimately lower the quality of testing. Dr López-Ríos recommends improving the communication within and between labs and understanding the limitations of the technologies (such as detection limits of EGFR testing, which are not always available in the package insert and would need to be determined by experimentation). For example, Dr López-Ríos comments that for real-time PCR the detection limit is not 1%, but closer to 3%.

With ALK testing, the controversy lies in using immunohistochemistry (IHC) or fluorescent in-situ hybridisation (FISH). Several studies have shown that the concordance rate using both technologies is high for both specificity and sensitivity. However, the use of a second assay in negative FISH cases is supported if the results are discordant, or the histology or clinical characteristics are challenging or different.

When testing for ROS-1 using IHC, if ROS-1+, then the result should be confirmed by FISH (or another molecular assay), likewise if testing using FISH, then consider confirming the result by IHC (or another molecular assay). Next-generation sequencing (NGS) is powerful and works well in most circumstances. In Dr López- Ríos’ lab, analysis of an NGS panel of 52 genes showed a 7.9% failure rate that was generally attributed to the RNA library preparation. This indicates that NGS should be coupled with ICH or FISH testing to confirm some results. There is a perception that NGS is the new gold standard, but sources of assay variation include the many available standard vendor solutions or customisation by the vendor depending on need. In addition, the technology itself can be different, e.g., amplicon-based or hybridisation capture-based. Using larger panels can result in less sensitivity and specificity, and panels can be incomplete (missing copy number variations or mutations, for example), causing samples to return a negative result.

Finally, in the analytical phase, Dr López-Ríos comments that we should consider testing for PDL-1 and CD-8 at the beginning of the pathology workflow.

Regarding interpretation, the challenge for Dr López-Ríos is in having too much information and being unsure what to do with it and what to report. They have created an intralaboratory molecular tumour board, where the reports (that integrate clinical, pathology and biomarker data), are reviewed prior to being sent out. They also correlate their findings with the published literature and discuss the need for additional tests, in the case of a bias in the bioinformatic algorithm for example, or in the rare case of a complete negative result.

Using case studies, Dr López-Ríos illustrates three points that demonstrate the importance of understanding the detection limit, molecular redundancy and rebiopsy in confirming initially negative results that were then positive upon reassessment.

In conclusion, to improve our approaches to targeted therapy using immunotherapy we need: a higher threshold in pre-analytics for tissue and blood-based approaches; to combine different methodologies in some—or even all—cases, which is not easy to do and involves many cost considerations; and to use an intralaboratory molecular tumour board that sits down to assess reports to help maximise the utility to the clinician and patients.

Question: Is cytological versus histological material still an issue? In 25% of cases, cytology is the only material available.
Answer: If doing cytology and you have a good specimen, put part of it into formalin for a cell block that can then be tested in clinical trials or labs.

Question: What is your percentage of patients with inadequate material at the time of diagnosis, where rebiopsy is necessary?
Answer: It would be nice to have greater involvement as a pathologist with the surgeon to collect adequate material, however this is not always possible. Generally the percentage of insufficient material is somewhere around 10-15%.