How Molecular Information Is Changing Diagnosis

As clinicians work to discover effective means to incorporate molecular information into clinical diagnoses, recent publications show that several distinct models are beginning to emerge. At the center of these efforts are innovative ways to better integrate complex molecular data with phenotypical presentation during the diagnostic process. DTET examines two models—one in oncology and one in genetics—that provide insight into how some institutions are working to accelerate the effective transition of sequencing-based testing into clinical care. In the first case, Moores Cancer Center (La Jolla, Calif.) describes its experience of establishing a molecular tumor board that incorporates the expertise of basic scientists and bioinformaticians into the clinical decision process. In the second case, German-based researchers have developed a software algorithm that can integrate targeted exome sequencing results with a structured set of phenotypic descriptors in order to increase the efficiency and scalability of medical genetic diagnoses. A molecular tumor board can provide multidisciplinary expertise that allows oncologists without advanced genomic training to incorporate molecular profiles of complex cancer care, according to a study published this summer in theOncologist. The molecular tumor board initiated at Moores included clinicians, basic scientists, geneticists, and bioinformatics/pathway scientists. Because molecular abnormalities “do not segregate by histology,” the experts represented multiple cancer types. The team was initiated in December 2012 and built on the “long-standing tradition” of oncology tumor boards. One-hour molecular tumor board meetings were held every two weeks with 25 to 40 people in attendance. Discussions focused on medical history, radiological findings, and pathology, including molecular profiling results. A consensus was reached for choice of agents most tailored to the patient’s specific aberrations. Molecular test results discussed by the panel included FoundationOne (Foundation Medicine; 33 tests), ResponseDx (Response Genetics; two tests) Molecular Intelligence (Caris Life Sciences; three tests), and full-exome sequencing (Champions Oncology; one test). At the time of publication, 34 patients’ cases had been presented representing breast cancer (n = 16), gastrointestinal cancers (n = 8), head and neck cancers (n = 4), and lung cancers (n = 2). Patients had received a median of three prior therapies in the metastatic setting, with prior or anticipated progression. The median total time from physician order to receipt of molecular diagnostic test results was 27 days. There was a median of four molecular aberrations (mutations, rearrangements, deletions, amplifications, or insertions) detected per patient, with a range of one to 14. A total of 74 genes were involved, with 123 distinct aberrations. The median time between receipt of molecular diagnostic results and the molecular tumor board presentation was 24 days. Three of the 11 evaluable patients treated on the basis of their molecular diagnostic results achieved a partial response (progression-free survival) despite having previously progressed on therapy. Four of the 11 patients had stable disease, while another four had progressive disease. The authors say the difficulty of treatment management informed by molecular diagnostics is witnessed by the nine patients who could not be treated because they were ineligible for an appropriately targeted clinical trial, logistically or financially could not enroll, or had inactionable aberrations. “This experience could be of significant importance to oncologists who are increasingly faced with advanced molecular diagnostic data, yet have minimal training in genomics,” write the authors, led by Maria Schwaederle, Pharm.D. “The attendance of basic scientists and bioinformatics and pathway analysts was crucial so that the therapeutic suggestions could be optimally informed by the results of molecular interrogation of the patients’ tumors.” In addition to the potential benefit of the individual patient, the molecular tumor board yielded broader benefits as well. The authors cite improved education to the treating physician and other attendees, as well improved workflow efficiency, including plans for new processes for tumor tissue acquisition and testing as a result of these meetings. Software Integrates Genomics, Phenotypical Analysis Integrating targeted next-generation sequencing (NGS) results into phenotype-driven bioinformatic analysis can provide quick and effective differential diagnostics in medical genetics, according to a study published Sept. 3 in Science Translational Medicine. The approach, the authors say, combines a structured set of phenotypic abnormalities with genetic variants to substantially improve the ranking of candidate genes, expediting clinical workflow. A traditional medical genetics evaluation inherently relies on recognizing a characteristic pattern of phenotypical presentations to guide targeted genetic testing for confirmation of the suspected diagnosis. However, diagnosis is complicated by the vast number of diseases and clinical syndromes, which may display in poorly understood phenotypic presentations along with high genetic heterogeneity. Even with extensive workups, fewer than half of patients with suspected genetic disorders are definitively diagnosed. While whole-exome sequencing (WES) is an increasingly promising option to end these diagnostic odysseys, challenges remain with interpretation, hampering the scalability of WES in clinical diagnostic settings. The Berlin-based researchers used variants in 2,741 of the most well-established Mendelian disease genes (the disease-associated genome [DAG]) to develop a targeted enrichment panel (7.1 Mb), which achieves a coverage of 20-fold or better for 98 percent of bases. The researchers then utilize an algorithm, the Phenotypic Interpretation of eXomes (PhenIX), that ranks variants based on pathogenicity and semantic similarity of patients’ phenotype (cataloged in the Human Phenotype Ontology set that uses more than 10,000 structured terms to describe 3,991 Mendelian diseases based on phenotypic abnormalities). The researchers found in computer simulations that relying on a variant score only ranks the true gene in first place less than 5 percent of the time, whereas PhenIX placed the correct gene in first place more than 86 percent of the time. The researchers applied PhenIX retrospectively to 52 patients with previously identified mutations (using Sanger sequencing) and known diagnoses. PhenIX achieved a mean 2.1 rank of the correct gene. In a prospective evaluation of PhenIX in 40 individuals without a diagnosis, PhenIX analysis enabled a diagnosis in 11 cases (28 percent; mean rank of 2.4). “On the basis of our results, we suggest that targeting all known disease genes, that is, a DAG, rather than the whole exome or genome, is advantageous in terms of target coverage, cost per sample, and the ability to provide quick and accurate clinical interpretation of the variants,” write the authors, led by Tomasz Zemojtel, Dr.Sc., Charité Universitätsmedizin Berlin. Among the benefits of this approach, the authors say, is that it allows a complete workup of targeted NGS results in roughly two hours per patient with a competitive diagnostic yield. Takeaway: Innovative methods, including molecular tumor boards and software algorithms, are being developed both in oncology and medical genetics in order to improve the efficiency and effectiveness of incorporating complex sequencing-based testing results into routine diagnosis and care.