Publication | Closed Access
A novel prognostic risk classification model for NUT midline carcinoma: a largest cohort analysis from the NMC registry.
14
Citations
0
References
2018
Year
Surgical OncologyEpidemiology Of CancerPrognosisPathologyLargest Cohort AnalysisCancer RegistrationSurvival Tree RegressionTumor BiologyPrognostic AssessmentOncologyBiostatisticsPublic HealthNmc RegistryMolecular OncologyCancer ResearchCancer PrognosisNut GeneNut Midline CarcinomaPrognostic BiomarkersCancer RiskCancer EpidemiologyMedicine
6085 Background: NUT midline carcinoma (NMC) is a rare subtype of squamous cancer defined by rearrangement of the NUT gene. Due to its rarity and under-diagnosis, there are no existing models to classify patients (pts) into risk groups based on baseline clinicopathologic factors. We aim to develop a prognostic risk classification model for NMC survival outcomes based on the largest cohort of NMC pts analyzed to date. Methods: Clinicopathologic variables and survival outcomes were extracted for N = 143 pts registered between 1990-2017 from the International NMC Registry. We performed survival tree regression to determine pt subgroups with statistically distinct risk factors and overall survival (OS) outcomes. Briefly, we performed Cox proportional-hazards regression for each potential factor. We dichotomized pts into two subgroups using the significant factor with the highest hazard ratio. We repeated this process within each subgroup until no further significant factors were found. Results: For N = 143 pts, median diagnosis age was 24 y (range = 18d-80y) and 48% were male. About half (54%) of tumors were without squamous cell differentiation; 54% had thoracic origin, 40% head/neck, and 6% other primary site. Most patients had the BRD4-NUT fusion (71%), followed by BRD3-NUT (13%), and NSD3-NUT (5%). At diagnosis, 78% had lymph node or organ metastases (mets). Median follow-up time was 2.9y (1d-19.1y). For N = 134 with survival data, median OS was 6.8m (95% CI = 5.8-9.7); 2-year OS was only 23% (±SE = ±4%). Survival tree regression identified 3 distinct risk groups: (A) no mets [2-yr OS = 39.5±10%; N = 24]; (B) with mets, non-thoracic origin [2-yr OS = 38.7±10%; N = 27]; (C) with mets, thoracic origin [2-yr OS = 6±4%, N = 55]. Conclusions: This is the first risk classification model for NMC. Metastatic pts with thoracic primary tumors have markedly poorer prognosis compared to other subgroups.