Sherrie Rice
October 2023—For molecular testing in oncology, the choice is often fast or slow. PCR-based platforms are rapid, and comprehensive genomic profiling by next-generation sequencing is slower, and each has its pros and cons.
At Massachusetts General Hospital, pathologists and others in the Center for Integrated Diagnostics in the Department of Pathology say they have the best of both for the treatment of non-small cell lung cancer.
For PCR-based platforms, turnaround time is measured in hours or days, and because the test is often for a single gene or a small number of targets, less tissue is required and complexity is lower. The large NGS-based panels look at hundreds of genes and can identify rare and multiple types of variants. TAT is measured in weeks.
The goal at the Center for Integrated Diagnostics “was for us to have our cake and eat it too,” said Lauren Ritterhouse, MD, PhD, then-associate director of the center and assistant professor, Harvard Medical School, in a CAP TODAY webinar made possible by a special educational grant from Thermo Fisher Scientific (www.captodayonline.com). (She is now senior laboratory director, Foundation Medicine.) The question for Dr. Ritterhouse and colleagues: “Can we have results rapidly, but can we also get more targets than just a couple of hotspots?”
The answer, they found, is yes. The rapid NGS test they are using now consists of 50 genes that cover the majority of the known actionable hotspots in lung cancer for which targeted therapies are available. The assay identifies more than 2,000 unique single nucleotide variants and indels, can detect more than 500 fusion-specific isoforms, and can call copy number alterations in 14 genes. “And the assay we use is an automated solution that facilitates a lot of the technical workflow in our laboratory, including nucleic acid purification and quantitation, the library preparation, sequencing, and data analysis,” Dr. Ritterhouse said.
The purification step is done in a separate instrument, and there are separate DNA and RNA extractions. The technology used for target enrichment is amplicon-based NGS. “While we think of it as a single test and it looks like a single test to our ordering providers, it’s two separate assays—DNA and RNA.”
The bioinformatics consists of molecular barcoding, targeted variant calling (de novo variant calling is not supported), tumor purity-adjusted CNV calling, and the ability to detect targeted isoform-specific fusions as well as non-targeted fusions, “so that we can detect an expression imbalance,” Dr. Ritterhouse said.
The laboratory validated the assay for formalin-fixed, paraffin-embedded tissue and frozen sections. “On average we use about five unstained sections for this testing platform, which is a minimal amount of tissue. And in our validation set, we were able to get down to about a millimeter square of tissue with successful results. So it works off a pretty small input,” she said.
The assay run time is about 16 to 18 hours. “Overall, end to end, the hands-on technologist time to run this assay is about two hours,” Dr. Ritterhouse said, including sample accessioning and label printing, for example. “The time spent on the assay instrument itself is quite a bit less.”
One of the technical supervisors in the lab did a comparative analysis of the resources the fast NGS and legacy methods require. “In trying to balance comprehensive genomic profiling and rapid turnaround time for quite a few years now, in our laboratory we pursued both in parallel,” she explained. “To try to get the best of both worlds, we ended up running quite a few assays”—perhaps a rapid fusion QPCR-based test for EGFR, for example, in addition to separate DNA-based and RNA-based comprehensive genomic profiling assays. “So all of these different tests would be run for a single case.”
The comparative analysis revealed the following: With rapid NGS, the cumulative hands-on time was reduced by 90 percent per sample, the time required for the associated training regimen and competency assessments was lower, total review time spent by attending pathologists fell almost 75 percent, and the up-front laboratory reagent costs dropped by more than 50 percent owing to less duplicative testing and fewer tests being run overall.
“We have over 50 different assays in our laboratory, ranging from PCR and FISH to NGS, and this is one of the easiest tests for us to train new technologists on because it is such an automated solution,” Dr. Ritterhouse said of the training time saved. The reduced number of tests run in the lab frees up time for the senior technologists to do valuable research, development, and validation work, she said. And the pathologist time saved is made possible by how the bioinformatics pipeline and variant calling are set up. “There is no de novo variant calling, so there’s a lot less time doing variant vetting and assessing annotations,” among other things. The rigorous pre-launch validation makes possible the attending pathologist’s “very expeditious review.”
Dr. Ritterhouse explained the workflow that makes it possible to get the results into the medical record the next day.
It begins with the technologist doing the laboratory information management system work—accessioning and label printing—and then the work to prepare the sample. If it’s FFPE tissue, it requires deparaffinization (20 minutes), preparing the lysis and PK tubes and then scraping the slides (five to 20 minutes), followed by two hours in incubation, during which the purification instrument is set up (30 minutes). Purification takes a few hours (12 samples at once maximum); while that runs, the sequencer prep and setup work is done (30 to 60 minutes).
Library prep, sequencing, and analysis are done overnight (12 to 16 hours). “We’ve built our own postanalytical data review interface and automated report generation. So that way, by mid-morning the next day, the data is available to look at and the reports are ready, in our lab information management system, for the attending pathologist to review before signing them out into the medical record.”
A text-based report is generated automatically; it pairs the variants based on tumor type and an internally developed variant database. It’s easily done, she said, because the pipeline calls only targeted SNVs and indels.
The center’s director, Jochen Lennerz, MD, PhD, and thoracic oncologist Ibiayi Dagogo-Jack, MD, spearheaded development of the ultra-rapid protocol used for lung cancer cases. The patient is scheduled for an interventional radiology biopsy, and the cytopathology fine-needle aspiration team performs the adequacy assessment. If it’s positive for malignancy, separate core biopsies are obtained, and it’s the core biopsies on which frozen sections are cut. “We usually get on average about 10 frozen sections from these core biopsies that are dedicated to the ultra-rapid molecular testing pathway. We have staff in our laboratory sharing a dedicated pager for this, and it goes off and says, ‘The sections are ready, come get them.’”
At the time of Dr. Ritterhouse’s webinar presentation in May, the laboratory had been running the assay for six months. “We’ve seen both in our validations and since our clinical launch really good performance,” she said. Clinical specificity is high: 99 percent for targeted SNVs and indels and 100 percent for fusion-specific isoforms. Clinical sensitivity is 95 percent for the targeted SNVs and indels. The assay’s limit of detection is five percent allele frequency, “which is great,” she said, “because it’s not infrequent that we get specimens with quite low tumor purity.”
The assay is targeted to run daily so results can be provided the next day.
The laboratory is open Monday through Friday only, so there are four to five runs weekly. NSCLC accounts for approximately 50 percent of the assay’s use, she said. To optimize cost efficiency, they want to run the assay at maximum capacity; thus sometimes other tumor types are included. The data she reported for five months of use showed, in part, that glioma accounted for five percent, melanoma and thyroid eight percent each, colorectal cancer six percent, head and neck and sarcoma five percent each, breast three percent. “And we’ve had good technical performance,” Dr. Ritterhouse said, “a less than one percent failure rate. It’s very common for our specimens to not have enough tissue, have poor DNA/RNA quality.”
In NSCLC, using the 50-gene focused NGS panel, driver alterations were identified in 75 percent of cases. “This includes several rare variants we’ve picked up that wouldn’t have been covered by our legacy single-gene or limited multiplex rapid assays,” she said. From the RNA component, the detection rate was 14 percent. “These are cases in which we’ve identified an ALK, ROS1, or RET fusion, or MET exon 14 skipping mutation.”
“So this is a powerful assay,” she continued, “in which for a large percentage of our non-small cell lung cancer cases, we’re able to get the results that our clinical colleagues need to treat the patients in a rapid, timely manner, with very good technical performance.”
Dr. Dagogo-Jack, assistant professor of medicine at Harvard Medical School and director of molecular integration, MGH Cancer Center, shared the cases of a few patients she had seen in clinic in the 10 days preceding the spring webinar.
The first was a 67-year-old woman with no smoking history and a new diagnosis of stage IIIA NSCLC. More recently, for a patient like this, “we’ve thought about chemotherapy plus immunotherapy up front,” she said. This patient’s tumor was sent for rapid NGS, and within 48 hours NGS revealed a KIF5B-RET fusion, which is not associated with sensitivity to immunotherapy.
“We were able to quickly avoid immunotherapy and recommend chemo-radiation therapy without delaying treatment,” Dr. Dagogo-Jack said.
The second case is an argument for keeping some of the legacy assays, she said, referring to EGFR, BRAF, and KRAS PCR assays. The patient was a 56-year-old woman with known KRAS G12C-mutant NSCLC who progressed on chemotherapy, immunotherapy, and a G12C inhibitor. She presented with a new contralateral pleural effusion.
“By all accounts, she shouldn’t develop a contralateral pleural effusion at resistance after undergoing a pleurodesis procedure,” Dr. Dagogo-Jack said. Her contralateral pleural effusion “is a bit perplexing because it’s not something we often see.” A thoracentesis was performed to confirm it’s the same cancer. “I want to know if a KRAS G12C mutation is present, and I want to do it quickly because she’s quite symptomatic. So here we use the KRAS PCR test, and we’re able to answer this question without getting more information than needed,” Dr. Dagogo-Jack said.
The third patient is a 66-year-old woman with no smoking history and stage III NSCLC that is abutting the aorta. “So there’s not a lot of room to make a mistake or pick the wrong therapy first,” she said. Initial tumor testing revealed ALK IHC positivity. “We need a reflex test that works quickly. FISH sometimes can take a few days.” They used the rapid fusion assay and were able to confirm an EML4-ALK arrangement. They avoided immunotherapy and proceeded with chemo-radiation therapy, “which we think is the best treatment for this condition. Control the cancer before it spreads or invades the aorta.”
What first steps to take to implement a rapid NGS testing program for NSCLC was a common question of the webinar attendees.
As for so many things, “a close working relationship across disciplines is central to the success of a program,” Dr. Dagogo-Jack said. Though it would be ideal to implement in every center, she acknowledged that many don’t have an in-house specialty pharmacy, as does MGH, or the necessary in-house diagnostic assay, particularly in a community hospital setting. “So I don’t think it’s feasible in all settings, but I hope the settings that are equipped with those resources might be able to implement similar programs.”
At MGH, she said, it took time to implement, and the process was iterative. “We started with proving we could do the rapid PCR test and, second, with proving that testing could be done on a fresh or frozen specimen. Only after that point and fine-tuning the process could we dial in the specialty pharmacy.”
But it all begins with conversation and education across specialties about the importance of adequate tissue and getting results quickly, she said. “It’s harder to implement if you don’t have the framework or foundation,” she added, by which other specialists are equally aware of the importance and invested in it.
Sherrie Rice is editor of CAP TODAY. View the full webinar online at www.captodayonline.com.