Associação entre os preditores mamográficos e ultrassonográficos do BI-RADS® com a agressividade biológica dos carcinomas de mama e com a sobrevida das pacientes

• Ano de defesa: 2024
• Autor(a) principal: Oliveira, Dênia Reis de Paula
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Câncer de mama; Mamografia; Ultrassonografia mamária; Radiogenômica; Sobrevida
• Link de acesso: https://hdl.handle.net/11449/255476

Objective: To evaluate the association between imaging characteristics of breast carcinomas (by the BI-RADS ® [5th Edition]) mammographic and ultrasonographic descriptors) and tumor histological characteristics, with the molecular subtypes approximated by immunohistochemistry (IHC) and with the overall survival (OS), Cancer Specific Survival (CCS) and Disease-Free Survival (DFS) of affected patients. Methods: This is a single-center, retrospective cohort. This study included patients diagnosed with invasive breast cancer between 2012 and 2015, age ≥ 18 years, in stages I, II and III. The included patients underwent diagnosis and follow-up at the Hospital das Clínicas de Botucatu, with mammography, ultrasound and histopathological data archived in the Hospital's medical records. Clinical, imaging and anatomopathological data were collected. The association between imaging descriptors, biomarkers and histological subtypes was analyzed. To study the association of variables and compare proportions, the chi-square test was used. The primary outcome was OS and secondary outcomes were DFS and CSS. Kaplan-Meier curve and Cox regression model were used to evaluate the association between imaging descriptors and OS, DFS and CSS. Differences in survival rates were assessed using hazard ratio (HR). In all tests, a significance level of 5% or the corresponding p-value was adopted. Results: The study included 260 women with a mean age of 56.3 ± 12.4 years at diagnosis, mean tumor size of de 2.6cm ±1,63 cm and follow up period of 114 months. The main mammography findings were: nodular lesions were associated with ER -, HER-2 negative and Ki-67 ≥ 20%. The irregular shape was associated with less aggressive lesions (RE+, RP+ and Ki-67 <20%). Round lesions were frequently related to Ki-67≥20%. Regarding the margin, while the circumscribed, microlobulated and obscured margins were linked to hormone receptor negativity, the spiculated margin was more frequent in ER and PR positive carcinomas. HER-2 suppression was more common in indistinct and microlobulated tumors, in the presence of asymmetries and suspicious microcalcifications. Finally, the suspicious axilla was more closely related to ER-, RP-, HER-2 + and high Ki-67. Regarding molecular subtypes, non-mass lesions and nodules with indistinct or microlobulated margins were associated with the hybrid subtype, while triple-negative (TN) tumors were mostly nodular, with an oval or round shape, with a circumscribed or obscured margin. The suspicious axilla in MMG was inversely associated with the Luminal B subtype and directly associated with pure HER-2. Also, in the HER-2 subtype, the presence of suspicious microcalcifications was statistically significant. Regarding ultrasound findings: of the microlobulated tumors, the majority were grade 3. Undifferentiated tumors were also more frequent with heterogeneous echogenicity, parallel orientation, posterior acoustic reinforcement and suspicious axillary lymph nodes. Grade 1 tumors were associated with hypoechogenic, non-parallel lesions, with shadow or combined pattern, with echogenic halo and negative axilla. Grade 2 carcinomas only showed an association with no subsequent effect. ER and PR positive tumors were irregular, with an angled, spiculated or indistinct margin, hypoechoic, non-parallel, with an echogenic halo, with shadow or absence of posterior acoustic effect and with a negative axilla. Negative ER and PR, in turn, presented an oval or round shape, microlobulated margins, heterogeneous echotexture, reinforcement, suspicious axilla on US and an inverse association with the presence of the halo. The mixed effect, despite being related to RE+, was mostly RP-. The PR was also negative in tumors that showed flow on Doppler. Regarding tumors with HER-2 overexpression, the minority had posterior shadowing. As for Ki-67, high rates were associated with oval or round lesions, with microlobulated or angled margins, heterogeneous, reinforcement, Doppler flow and suspicious axilla. Regarding subtypes, Luminal A were irregular, hypoechogenic, non-parallel, without Doppler flow or suspicious microcalcification, with posterior acoustic shadow and negative axilla. Luminal B, in turn, was associated with non-parallel orientation. HER-2 overexpressed tumors were associated with the presence of suspicious calcifications and suspicious axilla. TN tumors were mostly mass lesions, oval, with microlobulated margins, reinforcement, parallel orientation and absence of halo. Regarding survival outcomes, on MMG examination the presence of suspicious microcalcifications reduced OS and CSS rates. The type of shape and margin of the nodules modified the CSS and the axillary status decreased OS, CSS and DFS. On US, non-mass lesions resulted in lower CSS and the type of margin impacted OS and CSS. Heterogeneous lesions demonstrated lower CSS and DFS, while the posterior acoustic effect and axillary status had an impact on OS, CSS and DFS. Conclusion: The breast carcinomas imaging characteristics, evaluated by the mammographic and ultrasonographic BI-RADS ® (5th Edition) descriptors, in addition to being associated with biomarkers and histological subtypes approximated by IHC, also have prognostic value with a direct impact on overall survival, cancer-specific and disease-free survival.