CLN - The Sample

Automated Cervical Cancer Screening Equals Performance of Cytology Triage

The Sample: June 2018

A novel automated cervical cancer triage and screening strategy, incorporating a cytologic risk score algorithm and computer-interpreted cytology, performed comparably to traditional cytology triage and could bring high-quality cervical screening to underserved regions (J Natl Cancer Inst 2018;110:djy044). The findings “strongly support the feasibility of totally automated cervical screening without cytology,” according to the authors.

The novel approach identified 91.7% of high-risk human papillomavirus (HPV)-positive cases for immediate colposcopy, while discerning 38.4% of all HPV-positive cases recommended for retesting in 1 year. In contrast, a conventional strategy combining HPV typing and nonautomated cytology triage identified 89.1% of high-risk HPV-positive cases for immediate colposcopy and 37.4% for future retesting.

The novel method involves a slide scanner that captures images and features of slides, such as presence of different cell types, nuclear size, and nuclear contour. The algorithm generates a severity score and identifies the most innocuous slides in a batch to reduce and/or guide treatment decisions.

The study involved two parts, including developing and validating the risk score algorithm, and validating an association between higher risk scores and HPV-positive, abnormal cytology cotest results.

The investigators developed and validated the computer algorithm using residual cervical specimens and liquid-based cytology slides from the Kaiser Permanente Northern California (KPNC) cervical cancer screening program. They used 1,839 stored HPV test specimens and liquid-based cytology results from a subset of women who had had baseline or follow-up HPV-positive results.

To further validate the risk score algorithm, the researchers used a set of 243,807 slides from KPNC obtained during routine cervical screening in 2016 and 2017.

While the novel method performed comparably to the conventional strategy, the researchers noted that both methods are imperfect, with “not high” specificity, and with both identifying precancer cases the other missed.

The authors suggested that automated cervical screening and triage might be of particular interest in middle-resource settings because it would enable implementation of “high-quality cervical prevention programs” in areas that lack skilled cytology professionals. At the same time, labs in high-resource settings might want to adopt such a system as an alternative to conventional cytology practice.

Wide Variation in Lynch Syndrome Testing Practices

An international survey of institutions that specialize in research and clinical care involving Lynch syndrome found wide variations in management and testing practices (Clin Gastroenterol Hepatol 2018; doi:10.1016/j.cgh.2018.04.025).
The authors speculate that this could be due to rapid changes in testing technologies, differences in resources, and “lack of definitive data for many clinical questions.”

“In just a few decades, there has been a surge of advances in the knowledge and tools used for diagnosis and management of Lynch syndrome patients and families, and breakthroughs in the understanding and management of genetic cancer predisposition syndromes are accumulating quickly,” the authors noted.

The researchers conducted the survey of 128 members of the International Mismatch Repair Consortium, of which 49% from 21 countries responded, to assess potential targets for research and public policy efforts.

In terms of case-finding practices, 56% of respondents participate in automatic reflex testing of tumors, with 100% reporting testing of colorectal, 50% endometrial, and 13% ovarian cancers, respectively.

Nearly all (98%) respondents said they use immunohistochemistry when testing tumors, and three-quarters (78%) also perform microsatellite instability testing. Reported testing practices to distinguish sporadic mutations from Lynch syndrome cases included BRAF mutation (75%) and MLH1 promoter methylation (56%). Only about one-quarter (27%) said they use the new approach of testing for biallelic somatic mutations. Sites that do not perform the latter noted its relatively high cost in comparison to limited benefits.

Consistent with other studies, this survey found that less than 50% of at-risk family members had genetic testing.

Optimal Screening Strategy Could Reduce Progression to Multiple Myeloma by 19%

A computational model of progression from monoclonal gammopathy of undetermined significance (MGUS) to multiple myeloma (MM) suggests that starting screening for MGUS at age 55 and conducting follow-up screening every 6 years would reduce the prevalence of MM by 19% (Clin Cancer Inform 2018; Starting MGUS screening at age 65 and performing follow-up screening every 2 years would reduce prevalence by the same percentage.

The authors were interested in exploring optimal screening scenarios because of recent research findings suggesting that interventions like taking the antidiabetic agent metformin, and aspirin and losing weight, might slow or reduce progression of MGUS to MM. However, it is unclear how best to screen at-risk populations and how to assess the effect of these interventions at the population level.

The investigators developed a Markov chain model to depict the population dynamics of healthy individuals transitioning to undetected MGUS, detected MGUS, MM, and finally, death.

Their computation model considered life tables and epidemiologic data on MGUS and MM, which depend on genetic background, sex, and age, and correlate with ethnicity. The researchers fashioned screening scenarios based on a person’s age when first screened, the spacing between follow-up screens, and risk reduction after a positive screen.

In model simulations, they considered scenarios ranging from baseline low-risk MGUS incidence to elevated risk for certain groups, such as high-risk African-Americans, whose lifetime risk of MGUS is about twice that of individuals with low baseline risk.