External Quality Assessment Identifies Training Needs to Determine the Neoplastic Cell Content for Biomarker Testing

Abstract

Neoplastic cell content determination is crucial for biomarker testing. It is known that interobserver variation exists, but largescale data are missing about variation in tumor delineation and cell content determination. Results were obtained from the external quality assessment program for metastatic colorectal cancer from the European Society of Pathology (N = 5776 observations). The study included three parts: current practices were surveyed, neoplastic cell content estimations and delineations were retrieved from stained slides, and clinical reports were analyzed. Seventeen of 43 pathologists determined the neoplastic cell content in a tumor-rich area for DNA extraction and took immune cells (n = 37), tumor cell distribution (n = 33), desmoplastic stroma (n = 30), necrosis (n = 29), and mucus (n = 23) into account. The selected area was highly variable, and the average difference between the highest and lowest estimation ranged between 51% and 78% (2011 to 2017). The number of overestimations was alarmingly high in samples containing <30% tumor cells. Of concern is that 33 of 105 laboratories reported a wild-type result in a sample without tumor in 2017. Standardization of neoplastic cell content determination is needed for test outcome interpretation. The authors' data show variation in estimation practices, tumor delineations and estimations, and interpretation problems (n = 226 reports). Further training for selecting the most suitable block and creating clear reports is urgently needed.

Neoplastic cell percentage estimation in tissue samples for molecular oncology: recommendations from a modified Delphi study

“…Several studies have already shown significant variation in NCP estimates; on average, as high as 20% between different pathologists . This was recently confirmed by a large European‐wide study, which also identified large differences in the processes related to the estimation of the NCP, such as contouring the tumour area and macrodissection . In addition, this study identified problems with interpreting the NCP.…”

“…Accurate estimation is extremely important to correctly interpret the results of a molecular analysis which, in turn, is crucial for treatment decision‐making. Inaccurate estimations may impact upon patient outcomes . The present study proposes recommendations for standardising the processes related to defining the tumour area for macrodissection and for estimating the neoplastic cell percentage within this area.…”

“…For the first survey, statements regarding the determination of the neoplastic cell content were drafted based on literature review and a survey that was previously conducted as part of an external quality assessment scheme for colorectal cancer, as described elsewhere . The expert panel was asked to indicate the extent to which they agree with the statements and to explain their choice.…”

Managing Deviating EQA Results: A Survey to Assess the Corrective and Preventive Actions of Medical Laboratories Testing for Oncological Biomarkers

“…The fact that 13% of cases were denoted as insufficient materials is surprising, as the materials were carefully validated beforehand and verified to contain sufficient neoplastic cells lacking tumor heterogeneity, and other peers were able to successfully analyze these cases in the same scheme. Indeed, pre-analytical variables such as the estimation of the neoplastic cell content has previously been linked to deviations in mCRC EQA results [ 29 ]. As the samples in the EQA scheme consisted of pre-cut and labelled samples, laboratories should therefore be extra cautious as other pre-analytical issues could arise in routine practice during sample reception, storage, and preparation.…”

Recommendations for the clinical interpretation and reporting of copy number gains using gene panel NGS analysis in routine diagnostics

“…To relate the NGS detected gains to FISH detected gains, the calculated number of alleles could be corrected for the estimated percentage of neoplastic nuclei in the area from which the gDNA was isolated. While we realize the estimation of this percentage is error-prone [ 14 , 15 ], it can be supported by the variant allele frequencies (VAF) of somatic variants in other genes and it allows estimation of the number of gene copies in the “order of magnitude” required to asses clinical relevance. As mentioned above, the estimation of the actual copy number gains may be biased in highly aneuploid tumors.…”