Background: Systemic mastocytosis (SM), a rare, clonal mast cell disease associated with heterogenous clinical presentation, is characterized by severe, debilitating and often unpredictable skin, gastrointestinal, neurocognitive, and systemic symptoms (including life-threatening anaphylaxis), long-term co-morbidities, and in Advanced SM (AdvSM), organ damage due to mast cell infiltration. SM consists of 2 distinct groups of variants, advanced and non-advanced: AdvSM includes aggressive SM (ASM), SM with an associated hematologic neoplasm (SM-AHN), and mast cell leukemia (MCL); non-AdvSM includes indolent SM (ISM) and smoldering SM (SSM). Arriving at a definitive diagnosis may take years and is one of the challenges of SM management. To aid clinicians in the identification of SM, we developed and validated a diagnostic algorithm using data from select community hematology practices in the United States. Methods: A sample of 209 patients (SM and control) was obtained from the Quality Cancer Care Alliance (QCCA), a network of 19 community oncology and hematology practices. Data collection consisted of patient characteristics, laboratory parameters, and signs and symptoms at presentation. General linear models (GLM) with a logit link function were used in a backwards elimination process with the p value set at < 0.05 to identify patient factors at presentation that were associated with a diagnosis of SM. A Dx scoring algorithm (range: 0-26) was then derived from the final model coefficients, with the intent of correctly differentiating between SM and non-SM patients. A receiver operating characteristic (ROC) curve analysis was then done to measure the Dx accuracy of the algorithm. Results: Data from 105 SM and 104 non-SM control (diagnosed with hematological cancers) patients were collected from QCCA. The SM cohort included patients with different SM subtypes, including: ISM (47.6%); ASM (9.5%); SM-AHN (19.0%); MCL (1.9%); and subtype not documented (21.9%). The factors identified as being predictive ofa correct diagnosis of SM were 1) patient age, 2) lymph node status, 3) absolute neutrophil count, and 4-7) the following symptoms within 30 days of presentation: diarrhea, rash, skin lesions, and unintended weight loss. In the algorithm, patients were assigned scores for each diagnostic factor they possessed [Figure 1]. The final score was then associated with an overall percent likelihood of a positive SM diagnosis. The area under the ROC curve was 0.89 (95%CI 0.84-0.93), indicating good predictive accuracy. Patients with a total score of ≥ 9 units were found to be at highest risk for a positive SM diagnosis with the associated sensitivity, specificity, and positive predictive values of 84.8%, 76.0%, and 82.7%, respectively. Conclusions: This model indicates that the presence or absence of 7 characteristics may help determine whether patients are at increased risk of having SM. Use of this diagnostic algorithm may help facilitate earlier detection of SM by raising clinical suspicion to facilitate appropriate clinical workup, which would allow the timely initiation of effective targeted therapies for this rare disease. Additional application and external validation of the algorithm in a new sample of 162 patients managed through another community oncology network is ongoing.
Read full abstract