Real-world treatment patterns in patients with Sézary syndrome in the United States and the impact of Covid-19

Background: Sézary syndrome (SS) is an aggressive type of cutaneous T-cell lymphomas (CTCL). Due to its low prevalence, there are limited data on real-world treatment patterns of currently available SS therapies. Furthermore, recent approvals of new agents for patients with CTCL as well as COVID-19 likely impacted real-world treatment patterns. Objective: To examine real-world treatment patterns and the impact of COVID-19 among SS patients treated in 2018-2020 in the US. Methods: Patients with public or private insurance in the 2018-2020 Symphony Health Solutions database were classified into 3 groups: ≥1 diagnosis of SS (ICD-10-CM code: C84.1x) in 2018, 2019, and 2020, respectively. Patient characteristics and treatment patterns for all therapies recommended by the National Comprehensive Cancer Network guidelines version 2.2021 were examined: systemic therapy (e.g., extracorporeal photopheresis (ECP), parenteral, or oral agents), skin-directed therapy (SDT, e.g., topical, local radiation, total skin electron beam therapy, or phototherapy) and bone marrow transplant. The impact of COVID-19 was assessed via quarterly analysis. National drug codes, current procedural terminology and healthcare common procedure coding system codes were used to identify all treatments. Results: The analyses included 869, 882, and 853 SS patients in 2018, 2019, and 2020, respectively (mean age: 66.3, 66.9 and 67.3 years;male: 54.4%, 54.8%, and 55.6%). Overall, systemic therapy increased from 2018 to 2020 (41.8% to 46.5%), with increased parenteral (20.7% to 28.7%) but decreased ECP (17.0% to 13.5%) usage. SDT increased from 2018 to 2020 (48.9% to 52.9%), with increased topical (42.3% to 48.3%) but decreased phototherapy (6.3% to 4.1%) usage. ECP, mogamulizumab, and bexarotene were the most prescribed systemic therapies in 2019-2020, with mogamulizumab being the only one with increased usage over time. Quarterly analysis showed a decreasing ECP from Q1 to Q4 within each year, with a notable drop in Q2 2020. For parental systemics, there was an increasing trend in 2019 and 2020, but utilization in Q4 2020 was lower than that of Q3 2020. For oral systemic, there was a notable drop in Q2 2020 but an increased trend in Q3-Q4 2020. Conclusions: This claims analysis indicated increased use in systemic and SDT among SS patients in 2018-2020. The quarterly analysis indicated that the drop in ECP and oral systemic usage in Q2 2020 coincided with the onset of the pandemic, but there was a stable use of parenteral systemic during 2020.

A preliminary review of adoption of new HCPCS codes for NGS in CMS claims

Background: Adequate reimbursement is considered a prerequisite for adoption of new diagnostic technologies that facilitate patient access to better treatments. Detailed longitudinal investigation of the adoption of new HCPCS codes and the factors influencing it are scarce, although the availability of large-scale claims databases should facilitate such studies. We examined claims for three CPT codes used for next generation sequencing (NGS): 81445, 81450 and 81455 in a large database of claims data from CMS and attempted to correlate presumptive drivers of test adoption such as coverage decisions and payments with test volume. Methods: CMS claims data were accessed using CMS' Virtual Research Data Center (VRDC) under data use agreement 50486. Any claim with a CPT code of 81445, 81450 or 81455 was extracted from the data and analysed using SAS Enterprise Guide with results summarised in Microsoft Excel. Data relating to national/local coverage determinations were located by internet searches. Results: Test volumes for all 3 codes showed significant variability, including a large decrease around Q1-2 of 2020, likely due to the COVID-19 pandemic. Utilization of the 3 CPT codes varied by patient diagnosis. Details of the top 5 diagnoses for each CPT are given in the Table. The top 30 diagnoses for each CPT code accounted for 80.33%-88.45% of patients. Conclusions: Utilisation of NGS testing from 2016-2021 was highly variable, confounding attempts to match potential drivers to changes in monthly test volumes. A relatively small number of conditions accounted for >80% test use. Increased use of 81445 and 81450 from 2019 onwards may be related to CMS LCD issued in March 2018, suggesting that it can take 8-9 months or more for a LCD to filter through to testing practice. Decreases in test volume around March 2020 coincide with decreased patient presentation and testing for cancer because of the COVID-19 pandemic indicating that factors beyond reimbursement can significantly affect test use. Changes in reimbursement or adoption of proprietary lab analysis (PLA) codes covering specific NGS tests may have caused the drop in test volumes in the latter half of 2021. This study demonstrates that determination of factors affecting adoption of a test technology can be problematic due to wide variation in claims over a relatively short space of time. However, determination of these factors is important as they ultimately affect patient access to testing and potentially to therapy. (Table Presented).

Cardiovascular Computed Tomography And FFRCTIn The United States Medicare Population From 2013-2020

Introduction: Over the past decade, through numerous technical advances and clinical studies, cardiovascular computed tomography (CCT) has gained increasing acceptance;recently evidenced by receiving multiple class 1, level A recommendations in the 2021 AHA/ACC Chest Pain Guidelines. We aimed to evaluate recent CCT practice and practitioner trends in the US Medicare population with the motivation of guiding practice, training, and advocacy. Methods: A retrospective cross-sectional analysis of Medicare Part B pay-for-service physician payments was performed between 2013-2020. CCT/FFRCT exams and providers were identified by unique HCPCS codes. Providers, exams, cost, and payment denials were analyzed. Medical specialty, gender, and geo-location of providers were summarized. Results: From 2013 to 2019, the number of providers of CCT exams and the number of exams increased significantly. Providers of CAC scoring increased >210%. Providers of coronary CTA in the hospital setting increased 36% and in independent testing facilities by 9%. CAC scoring exams increased 724% and coronary CTA exams increased 126% (see Figure). In the first year of the COVID-19 pandemic (2020), CAC scoring usage decreased by -9.3% and coronary CTA by -3.3%. Since initial reimbursement in 2018, FFRCT usage has increased by 654% but was applied in only 4% of coronary CTA exams. In 2020, contrary to a moderate CCT exam decline, FFRCT analysis increased by 376% compared to the previous year. Medicare insurance acceptance of cardiac CT became more favorable into 2020 (see Figure). CAC scoring denials decreased from 61.6% to 33.2% and coronary CTA denials decreased slightly from 7.3% to 6.4%. FFRCT denials decreased significantly from 64% to 6%. In 2019, 30.5% of CCT providers were cardiologists with the remainder being predominantly radiologists. On the other hand, 76.2% of FFRCT providers were cardiologists. A slightly lower percentage of FFRCT billing physicians were female compared to CCT billing physicians (14.2% vs 17.9%). CA, NY, MN, TX, and PA had the highest FFRCT utilization. Conclusions: In general, both CAC scoring and coronary CTA utilization have increased, along with a large increase in the utilization of FFRCT over the study time period. This increase in utilization was accompanied by a significant increase in providers and a decrease in reimbursement denials. In the first year of the COVID-19 pandemic, CCT usage was robust and only decreased moderately. [Formula presented]

Categorizing Telemedicine Visits Using Natural Language Processing and Machine Learning

Objectives: Telemedicine visits increased recently due to the COVID-19 pandemic, however a discreet telemedicine indicator is not present in some Electronic Health Record (EHR) data. Natural Language Processing (NLP) and Machine Learning (ML) were to build a model to categorize patient visits to better understand telemedicine utilization. Methods: Initially, encounter type, note type, chief complaint, and appointment type were features used to categorize 389,315,647 visits spanning the last 14 years in an ambulatory EHR dataset. Each feature was filtered based on a list of 21 inclusion and 29 exclusion words or word chunks, as well as 7 CPT codes, 23 SNOMED codes, and 9 HCPCS codes. A clinician tagged each feature as indicating telemedicine or not. A predictive ML model was trained. Data was preprocessed by removing identifying features and punctuation, spelling correction, flagging negated words, and lemmatizing. Each feature was converted into unigrams, bigrams, and trigrams, and transformed with a TFIDF-vectorizer. The model was fit on a XGBoost ML model. The model tagged each feature as either 0 (not telemedicine) or 1 (telemedicine). Visits had multiple features that were conflicting. To determine if the whole visit was telemedicine or not, visit tie-breaking features were added if documented: vitals, labs, and medications prescribed. A rules-based model was created and applied to categorize the whole visit as telemedicine, not telemedicine, or not enough information. The visits were then re-evaluated by clinicians to determine overall model fitness. Results: In total, the model had an accuracy of 99.86% with an F1 score of 96.99%. There were 577% (3,450,561) more telemedicine visits since the start of the COVID-19 pandemic than the sum of all previous telemedicine visits combined (597,410). Conclusions: A mixture of ML and rules-based methods successfully categorize visits as telemedicine visits. Further studies stratifying on telemedicine and non-telemedicine visits can now be done.

Changes in management of TNBC durin the COVID19 pandemic of 2020

INTRODUCTION An increasing body of evidence demonstrates that the COVID-19 pandemic of 2020 saw large reductions in the number of US patients being diagnosed with a variety of conditions, including cancer. A previous real world evidence study based upon analysis of CMS claims data showed a large drop in cancer diagnoses across multiple solid tumor diseases and evidence suggesting changes in testing behaviors for these patients over the period of maximal lockdown measures to mitigate spread of infection. Further, the drop in patient numbers had not returned to normal once these measures were relaxed by the end of June. Therefore, we decided to examine CMS data for the entire year of 2020 and focus on a single sub-group in breast cancer, TNBC. These patients have poor prognosis and are relatively intensively managed;it was reasoned that changes in management, especially testing behavior, might be more apparent in this group than in breast cancer patients as a whole. METHODS CMS data for 2019-20 were queried using a proprietary business rule for identifying TNBC cases and then subdivided into 2 groups: those who received a treatment under a "J" HCPCS code and those who had not. Office visits, Level IV surgical pathology (SP) and immunohistochemistry (IHC) were defined by appropriate HCPCS codes. Since all PD-L1 testing is covered by HCPCS code 88360, a claim for 88360 was considered indicative of a PD-L1 test. A decrease in the number of patients during the COVID-19 pandemic Swas defined as a ≥ 10% drop for the value in a given month in 2020 compared to the same month in 2019, as a percentage of the 2019 median value. This is termed the "COVID-Dip". RESULTS Data were gathered from a total of 68, 018 patients, 8, 131 with a J code treatment and 59, 887 without. Results of COVID dip analysis are presented in Table 1. Trastuzumab administration showed an overall decline across the entire study period. While IHC for 88360 showed a COVID dip, administration of atezolizumab and pembrolizumab increased across the study period with administration of nivolumab (collectively immuno-oncology, IO, drugs) remaining relatively constant. 47% of patients receiving IO therapy received a presumed PD-L1 test. There was longitudinal variation in the use of chemotherapy agents but no apparent COVID dip in their use. DISCUSSION There were declines both in patient presentation to doctors' offices, as well as diagnostic testing among TNBC patients during the COVID-19 pandemic of 2020 with differences between those receiving chemotherapy under J codes and those not. There was no evidence of decline in use of chemotherapy under J codes. Increased IO use but declines in IHC testing suggest a greater use of off-label prescribing of these drugs during the pandemic. The decline in presentation to doctors' offices and in testing of patients not receiving J code drugs suggests that these patients may experience significant delays in management of their condition with concomitant increases in morbidity and mortality.

Real world evidence demonstrates replacement of 2D mammography with 3D mammography among insured women

Background Digital breast tomosynthesis (DBT), also called 3D mammography, was first approved by the Federal Drug Administration in 2011. The goal of 3D mammography is to improve accuracy compared to 2D digital mammography (DM), by increasing sensitivity and decreasing recall rates. To capture the broad utilization of DBT in populations receiving both screening and diagnostic imaging, this analysis investigates DBT usage over time in a longitudinal sample from 2016 through 2020 for adult women ranging from 18-74 years of age. Methods Retrospective analyses were conducted using de-identified administrative claims data from a large national U.S. health insurer. The study cohort consisted of women who were continuously enrolled in a commercial or Medicare Advantage plan from 1/2016 to 12/2020 and aged 18 to 74 years old as of 2016. All procedures were identified based on Current Procedural Terminology (CPT) and Healthcare Common Procedure Coding System (HCPCS) codes. For each study year, receipt of breast cancer screening with DM (S-DM) or including DBT (S-DBT) were captured. Receipt of DM and DBT not specified for screening based on CPT code descriptions were categorized as diagnostic, D-DM and D-DBT respectively. Women that received both DBT and DM in the same year were included in the DBT group. In addition, women who received MRI and ultrasound were also captured. Rates of each procedure by study year, insurance type, and age categories that align with recommended screening guidelines (<40 years old, 40-49 years old and 50-74 years old) were examined. Results Approximately 3.8 million women met study criteria;85% were commercially insured and 15% were Medicare Advantage. Table 1 shows rates of adult women who received mammography, MRI and ultrasound over the study period. About 74% of study subjects receiving screening were 50-74 years old at the start of the study period, 25% were 40-49 years and 1% were under 40 years old. In 2020, there were fewer women (3%-13%) receiving imaging procedures compared to counts in 2019. During the 5-year study period, there was a 3.5-fold increase in the number of women who received S-DBT. In 2016, 23% of women who received a screening mammogram received S-DBT and by 2020, this percent increased to 82%. The percent of women who received a diagnostic mammogram using D-DBT compared to D-DM also increased overtime;29% of women received a D-DBT in 2016 and this increased to 77% in 2020. The number of women with receipt of ultrasound and MRI were similar in each study year. The percent diagnostic/screening tests (including DBT and DM) were in the range of 18.5%-20.2% each year. Conclusion Among this cohort of women who were continuously enrolled in the health plan throughout the 5-year study period, this analysis shows that screening and diagnostic DBT utilization rates increased from 2016 to 2020 while DM screening and diagnostic imaging utilization concomitantly decreased. The percentages of women that received S-DBT and D-DBT were highest in 2020, even though 8%-13% fewer women had evidence of mammography than in 2019, which is largely due to COVID-19 related healthcare service disruptions. The rate of diagnostic tests as a percent of screening tests did not decrease with the adoption of DBT. Further analyses investigating rates of follow-up procedures and downstream costs are warranted.

Effect of COVID-19 Pandemic Control Measures on Diagnoses of Hematologic Cancer Among CMS Population in 2020

Introduction Measures taken to mitigate infection spread during the 2020 COVID-19 pandemic are considered to have caused significant unintended consequences on other diseases. Large decreases in the numbers of symptomatic and asymptomatic people presenting for diagnosis of heart disease, diabetes and cancer have been observed. A recent analysis of solid tumors showed up to 70% reduction in the number of patients presenting for diagnosis. The potential exists for significantly increased morbidity and mortality for these missed or delayed presenting patients. Further, it is important to determine whether infection spread mitigation measures affected the diagnostic testing and treatment decisions for these patients. This study aimed to determine whether pandemic control measures affected presentation, testing and treatment of patients across eight different hematologic cancers. Methods CMS claims data were analyzed for the presence of diagnostic (DX) ICD 10 codes indicative of hematologic cancer. Patients with a DX code first appearing in 2019 or in 2020 were selected to provide newly diagnosed pre-COVID-19 and during COVID-19 cohorts for comparison, with unique patient counts being calculated for each month. A “COVID-19 dip” i.e. a decrease in the number of patients was calculated as the change in number of patients diagnosed in a given month relative to the number for JAN2020. Dip duration was calculated only when the decrease was >10% of the JAN2020 figure. Patients who received treatment via a “J” code Healthcare Common Procedure Coding System (HCPCS) code were extracted from the cohorts and the time taken from initial diagnosis to first treatment calculated. Results Eight hematologic cancers: AML, CLL, CML, HEME (a group of different hematologic cancers), Hodgkins (HOG), Myelodysplasia (MDS), Non-Follicular Lymphomas (NFL), and Non-Hodgkins Lymphoma (NHL) showed a decrease in the number of patients being diagnosed during the early part of 2020 (Fig.1) Fig.1. Change in new patient diagnoses for selected hematologic cancers as a proportion of their JAN2020 value There was some variation in the depth and duration of the COVID-19 dip (Table 1) with MDS having both the longest and deepest dip. Median depth and duration of the dip was 33% and 3.5 months, respectively, with all dips starting either in FEB or MAR2020. Table 1. Duration and depth of COVID-19 dips for selected hematological cancers The proportions of patients receiving therapy via J HCPCS code (JRX) are shown in Table 2 Table 2. Proportions of patients receiving J code therapy Conclusions The decline in new patient diagnoses for heme cancers during the period when COVID-19 control measures were implemented is similar to that seen with solid tumors, although the depth of the COVID-19 dip was generally larger in the latter. There is no evidence of “catch up” diagnosis occurring i.e. patients missing from Q2 2020 are not reappearing en masse in subsequent quarters. The decline for MDS patients has, except for SEP to OCT2020, remained. Collectively, (depending on the calculation method), the COVID-19 dip for these eight heme cancers represents 16,584-33,671 patients who will likely have significantly increased rates of morbidity and mortality due to delayed diagnosis. Analysis of J code treatments show little difference between the proportions of patients receiving these treatments in 2020 compared to 2019 suggesting that at least some aspects of treatment e.g. infused chemotherapy, IO drugs for these patients was relatively unchanged by pandemic control measures. It also suggests that the main cause for decreased patient numbers treated is due to decreased testing for diagnosis, rather than not being treated once diagnosed. This aligns with findings from studies in the US and UK. The results of this study indicate that there may be a “backlog” of tens of thousands of people with cancer whose diagnosis has been significantly delayed and who urgently need to be identified in order to get on proper treatment to lessen the impact of that delay. [F rmula presented] Disclosures: No relevant conflicts of interest to declare.