Introduction: Men with lower urinary tract symptoms (LUTS) are heterogeneous, and treatment decisions are primarily based on reported symptoms, with varying degrees of success. Identification of clinically meaningful subtypes would allow for more personalized care. This study advances LURN phenotyping efforts by incorporating anatomic, metabolic, and inflammatory factors to previous phenotyping using urologic symptoms alone. Methods: 192 variables, including LUTS, bladder diary, demographic, medical history, and physical exam data from LURN were assessed among 519 men with at least one bothersome urinary symptom, using weighted Tanimoto indices, semi-supervised learning, and resampling based consensus clustering, to identify clusters of participants. Differentially abundant proteins (among 276 cardio-metabolic, inflammatory, and neurologic proteins) from serum of 220 of these men were compared across the identified clusters. Results: Five clusters of men with LUTS were identified (figure). Two clusters reported mild LUTS (refined male cluster [RM]1: n=66 and RM2: n=84). RM1 was older than RM2 (70 vs 56 years), had more comorbidities (FCI 2.4 vs 1.5) and erectile dysfunction (IIEF: 1.8 vs 4.3). Two BPH-like clusters were identified (RM3: n=64 and RM4: n=188), with voiding and post micturition symptoms. RM3 resembled classic BPH and had a large PVR (275ml vs 46ml), while RM4 reported more frequency, urgency, incontinence, pain, and psychosocial symptoms (PROMIS depression 51 vs 46, PROMIS anxiety 52 vs 46, PSS 14 vs 8), resembling BPH with OAB. RM5 (n=119) was characterized by OAB with urgency, incontinence, and frequency. RM5 had more comorbidities (FCI=3) and psychosocial symptoms (PSS=12). Targeted proteomics analysis corroborated the five clusters. Between 15 (RM2) and 87 (RM1) differentially abundant proteins were identified within each cluster. Minimal overlap was observed across proteins and affected pathways within the clusters. Conclusions: We discovered 5 clusters of men with LUTS. The minimal overlap of differentially abundant proteins across clusters suggest that identified subtypes are biochemically distinct. Subtypes need to be clinically validated but may represent populations with distinct pathophysiology and therapeutic needs. SOURCE OF Funding: NIH/NIDDK
