In an effort to improve cancer diagnosis, the National Academy of Medicine (NAM)-part of the National Academies of Sciences, Engineering, and Medicine-held a comprehensive virtual workshop to examine diagnosis from many perspectives. The webinar, held with support from the Gordon and Betty Moore Foundation, is part of a NAM diagnostic series including sepsis and acute cardiovascular disease.
The workshop covered diagnostic pathways for screening and symptom-detected cancers, new technologies to improve diagnosis, strategies to mitigate population disparities in treatment outcomes, the need to include the patient voice, best practices in cancer screening and early detection, and creating a vision for excellence in cancer diagnosis.
The NAM has explored this topic in previous reports, including "Delivering High-Quality Cancer Care: Charting a New Course for a System in Crisis." That 2013 report stated, "The complexity of cancer impedes the ability of clinicians, patients, and their families to formulate plans of care with the necessary speed, precision, and quality." It recommended a conceptual framework which envisions a learning health care information technology system that "supports patient-clinician interactions by providing patients and clinicians with the information and tools necessary to make well-informed medical decisions."
In cancer, "the timeline for diagnosis is often stretched out," noted Andrew Bindman, MD, Planning Committee Chair of the workshop, who is Executive Vice President and Chief Medical Officer for the Kaiser Foundation Health Plan, Inc., and Hospitals. Cancer often presents in a primary care office, and the timeline itself gives opportunities for improvement, he said.
Bindman explained that diagnosis is a process, and it isn't always resolved before treatment decisions need to be made. He noted that the diagnostic process can be fraught with siloed specialists, and stressed the importance of listening to patients' and caregivers' voices to achieve an accurate and timely cancer diagnosis. Bindman said the process of cancer diagnosis works much better for patients if a physician makes a diagnosis "with them rather than at them."
Missed and delayed cancer diagnoses are common and consequential, and are the number one cause of medical malpractice cases, said Daniel Yang, MD, Program Director at the Gordon and Betty Moore Foundation. "We think diagnosis should be safe, timely, effective, efficient, equitable, and patient-centered," he emphasized.
Diagnostic delays are the most expensive and tragic malpractice cases, and about half of these are cancer misdiagnoses, said Luke Sato, MD, Chief Medical Officer and Senior Vice President at CRICO/RMF (Risk Management Foundation of the Harvard Medical Institutions) and Assistant Professor of Medicine at Harvard Medical School.
"The time to first treatment is an outcome that we can really work toward" in terms of improving the diagnostic process, said Mia Levy, MD, PhD, Chief Medical Officer for Foundation Medicine and former Director of the Cancer Center at Rush University Medical Center.
Levy said that what she and her colleagues see is a lot of system barriers in getting to that first cancer treatment. "Insurance authorization is a big one," she said, noting that insurance companies must be held accountable for their role in timely cancer diagnosis. Two cancer survivors/patient advocates, Tamika Felder and Gwen Darien, emphasized just how stressful it is for patients to wait for a cancer diagnosis and for their first treatment to start.
While there are screening tests for the big four cancers-lung, colorectal, breast, and prostate-speakers noted that these tests are underused (especially in disadvantaged populations), most cancers do not have screening tests, and most cancers are diagnosed because of symptoms.
"Screening is not available for most cancer types," said Chyke Doubeni, MD, MPH, Professor of Family Medicine, founding Director of the Mayo Clinic Center for Health Equity and Community Engagement Research, Associate Director for Community Engagement in the Mayo Clinic Cancer Center, and a member of the U.S. Preventive Services Task Force.
Doubeni noted that screening is a multistep process, and when a cancer screening test exists, for the screening process to be successful it should include age-appropriate, timely screening intervals, access to high-quality screening, timely follow-up after an abnormal finding, and timely delivery of treatment when a diagnosis is confirmed. Cancer is currently screened one at a time, and Doubeni was asked if there could be a way to bundle tests to detect multiple cancers. While that is the ideal, "unfortunately we're not there yet," he said.
From the primary care perspective, it is not easy to diagnose a cancer, noted Alex Krist, MD, MPH, Professor of Family Medicine and Population Health at Virginia Commonwealth University, and a practicing family physician and resident educator at the Fairfax Family Practice Residency. For example, he said that, in his community, many oncologists will not see a patient until she or he has a tissue diagnosis-which is often difficult for a primary care practitioner to obtain.
"Most cancers are detected after symptomatic presentation," said Elizabeth Sarma, PhD, MPH, Program Director in the National Cancer Institute's Health Systems and Interventions Research Branch of the Healthcare Delivery Research Program.
Recognizing that reality-and the fact that the United Kingdom (UK) has lagged behind similar countries in cancer diagnosis-the UK undertook an initiative to increase the detection of early-stage cancers from 50 percent to 75 percent, said Fiona Walter, MD, MA, Co-Leader of the Cancer Research UK-funded multi-institutional, multi-national CanTest Collaborative, whose goal is to accelerate progress toward improving cancer outcomes by focusing on the transformative implementation in primary care of tests to support the early detection of cancer.
Walter is Professor of Primary Care Cancer Research and Director of the Wolfson Institute of Population Health at Queen Mary University in London. In primary care in the UK, few biomarker or imaging tests have traditionally been available in primary care, said Walter, but that is changing. She described how the UK is using decision-support tools that predict risk thresholds for constellations of symptoms to decide whether to further investigate a patient's symptoms or refer that patient. The risk threshold tools were developed through algorithms that predict individual risk based on symptoms and other factors. Walter said one recent finding in the UK is that the cancer antigen biomarker blood test CA-125 is more predictive of ovarian cancer in women over age 50 than previously thought.
Technologies to Enhance Diagnosis
The NAM webinar showcased some promising technologies that have much potential to improve and enhance cancer diagnosis. These include the following:
* A liquid biopsy pre-screening tool for cancer detection. This novel blood test employs DELFI (DNA evaluation of fragments for early interception), an algorithm-based technique that spots unique "packaging" patterns in the fragmentation of DNA shed from cancer cells circulating in the bloodstream. DELFI, which is being studied in clinical trials, is highly sensitive and specific across seven cancer types, said Victor E. Velculescu, MD, PhD, Professor of Oncology and Co-Director of the Cancer Biology Program at the Johns Hopkins Sidney Kimmel Cancer Center. He noted that, since primary care physicians are on the front lines of cancer diagnosis, the DELFI liquid biopsy test "could be very helpful to them."
* The use of artificial intelligence (AI) to improve diagnosis via imaging. AI can be applied in precision oncology as a computer aid to a radiologist and as a second reader, said Maryellen Giger, PhD, the A.N. Pritzker Distinguished Service Professor of Radiology, Committee on Medical Physics, and the College at the University of Chicago. "We need a representative dataset" using AI in imaging, stressed Giger. She suggested that the AI community embrace data commons in order to increase and enhance the use of ethically sourced AI tools.
* The use of AI to improve cancer diagnosis via pathology. "The potential of AI-driven pathology is very broad," said Andrew H. Beck, MD, PhD, co-founder of PathAI and former faculty member in the Department of Pathology at Harvard Medical School. He noted that diagnostic variability-different diagnoses from different pathologists on the same biopsy-"has been a problem in pathology for decades." Beck said AI deep learning "has shown remarkable progress in recent years," and this technology is developing into an accurate diagnostic platform for pathologists.
* Molecular/genetic tumor characterization in precision oncology care. Tumor profiling is leading to improved diagnosis and thus improved treatment through targeted drugs, said James M. Ford, MD, Professor of Medicine (Oncology) and Genetics; Director of the Stanford Cancer Genetics Clinic; and Director of the Cancer Genomics Program at Stanford University. "The future is bright" for precision cancer diagnosis, and he noted that many cancer centers now have molecular tumor boards, citing as an example of precision oncology STAMP (Stanford Solid Tumor Actionable Mutation Panel). Ford said the challenge now is whom to test and at what stage, taking into account tumor heterogeneity, testing costs, and access issues.
Meeting Takeaways
In summing up take-home messages from the day-long webinar, Bindman cited the following.
* Waiting for a cancer diagnosis is stressful and nerve-wracking for patients, and health care providers need to work as quickly as possible to reach an accurate diagnosis.
* Time is of the essence in cancer diagnosis in order to begin the correct treatment as early as possible to achieve the best outcome.
* The UK experience shows that symptom assessment is vitally important, and depending on screening-even if screening tests exist for certain cancers-is not enough. Listening to the patient's voice in symptom assessment is essential.
* History teaches that the use of new technology-beneficial though it may be-can exacerbate health inequities. So in using proven new diagnostic tools, health professionals must be cautious and careful to make sure they are benefiting all cancer patients.
* Errors in cancer diagnosis persist because of a lack of coordinated, comprehensive care systems-which the COVID-19 pandemic has exposed. Every effort should be made to avoid fragmented, siloed care. "We can do better with the tools we have," Bindman noted.
Peggy Eastman is a contributing writer.