RESUMEN
BACKGROUND: Repeat leiomyoma occurrence or even reintervention is common after myomectomy. Little is known about the factors related to repeat interventions. OBJECTIVE: This study aimed to determine the frequency of leiomyoma-related reintervention after an initial laparoscopic or abdominal myomectomy and to analyze both clinical and molecular risk factors for reinterventions. Another objective was to define the frequency of clonally related tumors from repeat operations. STUDY DESIGN: This retrospective cohort study included 234 women who had undergone laparoscopic or abdominal myomectomy in 2009 to 2014. Information on repeat leiomyoma-related interventions as well as on other clinical factors was collected from medical records after a median follow-up time of 11.4 years (range 7.9-13.8 years) after the index procedure. The effect of clinical risk factors on the risk of reintervention was analyzed by the Kaplan-Meier estimator and the Cox proportional hazards model. For molecular analyses, we examined the mutation profiles of 133 formalin-fixed paraffin-embedded leiomyoma samples from 33 patients with repeat operations. We screened the tumors for the 3 primary leiomyoma driver alterations-mediator complex subunit 12 mutations, high mobility group AT-hook 2 overexpression, and fumarate hydratase-deficiency-utilizing Sanger sequencing and immunohistochemistry. To further assess the clonal relationship of the tumors, we executed whole-exome sequencing for 52 leiomyomas from 21 patients who exhibited the same driver alteration in tumors obtained from multiple procedures. RESULTS: Reintervention rate at 11.4 years after myomectomy was 20% (46/234). Number of leiomyomas removed at the index myomectomy was a risk factor (hazard ratio 1.21; 95% confidence interval 1.09-1.34). Age at index myomectomy (hazard ratio 0.94; 95% confidence interval 0.89-0.99) and postoperative parity (hazard ratio 0.23; 95% confidence interval 0.09-0.60) were protective factors. Molecular characterization of tumors from index and nonindex operations confirmed a clonal relationship of the tumors in 3/33 (9%) patients. None of the leiomyomas harboring a mediator complex subunit 12 mutation-the most common leiomyoma driver-were confirmed clonally related. Fumarate hydratase-deficiency was detected in repeat leiomyomas from 3/33 (9%) patients. All these patients harbored a germline fumarate hydratase mutation, which is distinctive for the hereditary leiomyomatosis and renal cell cancer syndrome. Finally, we identified 3 (3/33; 9%) patients with 2 tumors each displaying somatic mutations in a recently identified novel leiomyoma driver gene, YEATS domain-containing protein 4. All YEATS domain-containing protein 4 mutations were different and thus the tumors were not clonally related. CONCLUSION: Our study shows that reintervention is common after surgical myomectomy. Uterine leiomyomas typically develop independently, but some share a clonal origin. Repeat leiomyoma occurrence may be due to genetic predisposition, such as a germline fumarate hydratase mutation. Distinct somatic YEATS domain-containing protein 4 mutations identified in multiple leiomyomas from the same patient indicate a possible role for YEATS domain-containing protein 4 in repeat leiomyomas.
RESUMEN
BACKGROUND: Biological heterogeneity of large B cell lymphomas (LBCLs) is poorly captured by current prognostic tools, hampering optimal treatment decisions. METHODS: We dissected the levels of 1,463 serum proteins in a uniformly treated trial cohort of 109 patients with high-risk primary LBCL (ClinicalTrials.gov: NCT01325194) and correlated the profiles with molecular data from tumor tissue and circulating tumor DNA (ctDNA) together with clinical data. FINDINGS: We discovered clinically and biologically relevant associations beyond established clinical estimates and ctDNA. We identified an inflamed serum protein profile, which reflected host response to lymphoma, associated with inflamed and exhausted tumor microenvironment features and high ctDNA burden, and translated to poor outcome. We composed an inflammation score based on the identified inflammatory proteins and used the score to predict survival in an independent LBCL trial cohort (ClinicalTrials.gov: NCT03293173). Furthermore, joint analyses with ctDNA uncovered multiple serum proteins that correlate with tumor burden. We found that SERPINA9, TACI, and TARC complement minimally invasive subtype profiling and that TACI and TARC can be used to evaluate treatment response in a subtype-dependent manner in the liquid biopsy. CONCLUSIONS: Altogether, we discovered distinct serum protein landscapes that dissect the heterogeneity of LBCLs and provide agile, minimally invasive tools for precision oncology. FUNDING: This research was funded by grants from the Research Council of Finland, Finnish Cancer Organizations, Sigrid Juselius Foundation, University of Helsinki, iCAN Digital Precision Cancer Medicine Flagship, Orion Research Foundation sr, and Helsinki University Hospital.