Refinement of Hyams Grading Criteria for Olfactory Neuroblastoma in an Endoscopic Sinus Surgery Predominant Cohort with Extended Follow-up: Worth the Effort?

Background: Hyams grading is considered prognostic in olfactory neuroblastoma (ONB), but grading criteria are subjective and predate modern classification and surgical approach. We evaluate the application of granular grading criteria to an endoscopic surgery predominant cohort with extended follow-up.

Methods: 78 ONB patients were identified (1994-2019) with original diagnoses dating to 1979. Original Hyams grade, Modified Hyams grade incorporating more detailed criteria and other histologic features (i.e. divergent differentiation, clear cell/oncocytic change, and spindling) were assessed if feasible, evaluated for distribution by grade and correlated with outcomes.

Results: Original Hyams grade (n = 43) distribution was: I: 4 (9%), II: 27 (63%), III: 11 (26%), and IV: 1 (2%). Modified Hyams grade (n = 59) distribution was: II: 29 (49%), III: 26 (44%), and IV: 4 (7%) with no grade I cases. Cases were more frequently upgraded with Modified Hyams grade. Mitotic rate (adjusted p < 0.001), pleomorphism (p < 0.001) and absence of rosettes (p = 0.002) were the only features that varied significantly between low grade (I-II) and high grade (III-IV) ONB. Neither Original nor Modified Hyams grade correlated significantly with any outcome endpoints by univariable Cox regression analysis (Median follow-up on surviving patients: 121.8 months). Splitting Modified Hyams grade into low and high trended towards a significant association with disease free survival on Kaplan Meier analysis (DFS, log rank p = 0.088). Severe nuclear pleomorphism correlated with adverse disease specific survival (Hazard ratio (HR): 8.092, p = 0.024) and DFS (HR:3.81, p = 0.033). On multivariable Cox regression, only procedure type (combined and transcranial approaches) remained significant prognosticators [HR:8.616, p = 0.004, and HR: 17.559, p = 0.017, respectively] with Kadish-Morita Stage D being nearly significant [HR:14.35, p = 0.054].

Conclusion: Modified Hyams grade offers a slight improvement over Original Hyams grade but is still not particularly impactful as compared to disease extent/stage and procedure type. Hyams grade I and IV are rare and likely unnecessary, justifying resolution of ONB grading to simply high and low grade. Furthermore, it appears that only a subset of parameters for Hyams grade is relevant in terms of classification and outcome.