Summary
Clinicians always want to do what is right for their patient, and that includes keeping invasive diagnostic procedures to a minimum. But for those diagnosing cutaneous melanoma, ordering biopsies for lesions that are ultimately benign is par for the course. So much so that one of the key diagnostic tools is the number-needed-to-biopsy (NNB) metric, which provides clinicians a general sense of the number of benign biopsies performed.
To explore additional resources offered by Oncology Times and to subscribe, visit https://journals.lww.com/oncology-times/pages/default.aspx
Original Article
A Call for Melanoma Biopsy Standards
Oncology Times
By Rebecca Hepp
Clinicians always want to do what is right for their patient, and that includes keeping invasive diagnostic procedures to a minimum. But for those diagnosing cutaneous melanoma, ordering biopsies for lesions that are ultimately benign is par for the course. So much so that one of the key diagnostic tools is the number-needed-to-biopsy (NNB) metric, which provides clinicians a general sense of the number of benign biopsies performed.
“The NNB for cutaneous melanoma is an imperfect metric, as it captures more than just the provider’s diagnostic accuracy, or ability to distinguish benign from malignant,” according to Kelly Nelson, MD, FAAD, Associate Professor in the Department of Dermatology, Melanoma and Skin Center, at The University of Texas MD Anderson Cancer Center. Many factors influence the NNB, she explained, including disease prevalence in a provider’s patient population. More melanomas within a population mean it will take fewer biopsies to find them, while fewer melanomas within a population leads to a higher NNB. Thus, this number is a moving target, according to her new meta-analysis.
“Currently, there are no standardized metrics for provider skin cancer screening,” Nelson admitted. To find out what information is available, she and her team set out to consolidate the publications addressing the NNB for cutaneous melanoma and quantify the breadth of the published literature, she explained.
“It's important to emphasize that our research does not assert a specific ‘ideal’ NNB, as removal of benign nevi can be appropriate and necessary,” Nelson clarified. “Instead, we are starting a conversation about the metric and how we can develop appropriate, quality metrics for melanoma screening.”
Study Details
To do this, she and her team pooled data from 46 studies over an 8-year span and found the NNB for melanoma diagnosis varies substantially and is affected by everything from clinician specialization and the use of dermoscopy to location-specific melanoma prevalence and even patient-specific factors such as melanoma subtype and patient anxiety.
Their analysis included at least 3,681 clinicians and 455,496 biopsied tumors, most of which were pulled from retrospective medical record reviews. The number of clinicians may be even higher, the team noted, considering some of the studies did not specifically address the number of clinicians included in the analysis. The researchers broke the data into three groups: all biopsied tumors (stratum A), all melanocytic tumors on pathologic review (stratum B), and tumors of clinical concern for melanoma (stratum C). They found 12 percent of biopsies were identified as cancerous within each of the three strata, with a total of 29,257 cutaneous melanomas. The overall weighted mean NNB came in at 15.6, with a range of 2.2 to 287. Once the team removed stratum A from the equation, that average NNB dropped to 14.8 for all clinicians.
When the researchers looked at NNB based on clinician specialty, they noted significant differences. Dermatologists worldwide had a mean NNB of 7.5, while Australian primary care physicians (PCP)—the only PCP group specifically identified in the studies—came in at 14.6 and U.S.-based clinicians had a mean NNB of 13.2.
Nelson was surprised to find NNB varied significantly even within a single institution, suggesting clinician-specific variables at play. In dermatology in particular, older age, residency training, and using dermoscopy and total body photography all produced lower NNB.
“When we look at providers in the same practice, assuming the frequency of melanoma is roughly similar across their patients, different providers have different NNB, suggesting that provider-directed educational efforts could support improvements in NNB,” she said.
This is where her new program, Project ECHO (Extension for Community Healthcare Outcomes), may help. Using didactic lectures through a telemonitoring platform, the program helps clinicians develop new clinical skills such as dermoscopy—the proficient use of which improves NNB statistics, according to Nelson’s study and a recent Cochrane Review (2018; doi: 10.1002/14651858.CD011902.pub2). Nelson, along with Stephanie Savory, MD, Assistant Professor of Dermatology, UT Southwestern Medical Center, and Janice Wilson, MD, Assistant Professor of Dermatology, University of Texas Medical Branch, is using Project ECHO to teach dermoscopy to multiple dermatology residency programs across Texas, and Nelson hopes to expand their efforts to primary care providers in the near future.
Research Hurdles
The meta-analysis, while incredibly insightful, was also an eye-opener into the lack of research and standardization. Many studies didn’t include the number of participating clinicians, clinician specialization and training, use of dermoscopy, inclusion or exclusion criteria, study methods and even reporting of total tumor counts.
The data provided a decent real-world picture of clinical practice overall, according to the researchers, but the lack of U.S.-specific NNB data is concerning. Only one study from Australia specified PCP data, and that country is well-known for its higher rate of melanoma and culture of PCP-based screening, the study said. The findings likely do not reflect U.S.-based practices, the researchers noted.
Another key concern with the data is the inconsistency of reporting the degree of clinician concern for melanoma. The team suspects some clinicians may classify tumors as concerning, whether or not they truly are, to ensure insurance covers the biopsy. The data also lacked information on false-negatives, which could have significant ramifications for cost and cost-effectiveness analyses.
“The wide range of NNB, paucity of U.S.-based clinician data, and varying study designs reported in the included publications highlight the potential advantage of developing consensus-led, structured clinical impression and pathologic diagnostic data to standardize NNB reporting across electronic health records and health care systems,” the authors concluded.
Improving Clinical Practice
Given the inherent variability in NNB reporting, incorporating the NNB into clinical practice can be challenging without the ability to create custom data reports, Nelson explained. To do that, clinicians need access to more user-friendly reporting tools so they can better document their clinical impression and diagnostic data. This, along with more stringent research standards, can go a long way to streamline NNB reporting and provide a more accurate picture. Nelson and her team are already working on creating structured data fields for their own electronic health record, which she hopes will help facilitate accurate NNB reporting in the future.
Rebecca Hepp is a contributing writer.