Use of artificial intelligence to predict response to neoadjuvant chemotherapy in breast cancer

:Breast cancer is the object of thousands of studies worldwide. Nevertheless, few tools are available to corroborate prediction of response to neoadjuvant chemotherapy. Artificial intelligence is being researched for its potential utility in several fields of knowledge, including oncology. The development of a standardized Artificial intelligence-based predictive model for patients with breast cancer may help make clinical management more personalized and effective. We aimed to apply Artificial intelligence models to predict the response to neoadjuvant chemotherapy based solely on clinical and pathological data. Methods: Medical records of 130 patients treated with neoadjuvant chemotherapy were reviewed and divided into two groups: 90 samples to train the network and 40 samples to perform prospective testingand validate the results obtained by the Artificial intelligence method. Results: Using clinicopathologic data alone, the artificial neural network was able to correctly predict pathologic complete response in 83.3% of the cases. It also correctly predicted 95.6% of locoregional recurrence, as well as correctly determined whether patients were alive or dead at a given time point in 90% of the time. To date, no published research has used clinicopathologic data to predict the response to neoadjuvant chemotherapy in patients with breast cancer, thus highlighting the importance of the present study. Conclusions: Artificial neural network may become an interesting tool for predicting response to neoadjuvant chemotherapy, locoregional recurrence, systemic disease progression, and survival in patients with breast cancer (AU)

Evaluation of clinical and pathological response factors to neoadjuvant chemotherapy in breast cancer patients

Objectives: To evaluate breast cancer (BC) patients treated with neoadjuvant chemotherapy (NACT) and to analyze clinicopathological features correlating with pathological complete response (PCR) and survival outcomes. Methods: Observational, descriptive, and retrospective study. The medical records of BC patients who underwent NACT were reviewed and analyzed using the Statistical Package for the Social Sciences (SPSS), version 20.0. Results: Of the 176 BC patints who underwent NACT, 62 patients (35.2%) achieved PCR. The PCR rate was 22% (n = 2) for luminal A, 15% (n = 9) for luminal B/HER2-negative, 45.5% (n = 15) for luminal B/ HER2-positive, 50% (n = 14) for non-luminal/HER2-positive, and 47.8% (n = 22) for triple-negative (p = 0.01). Histological grade, estrogen receptor (ER) expression, progesterone receptor (PR) expression, and HER2 status were significantly associated with PCR (p = 0.022, p = 0.01, p = 0.01, and p = 0.02, respectively). The median follow-up was 35.9 months, the estimated 5-year disease-free survival (DFS) was 96.7% in the PCR group and 83.2% in the non-PCR group (p = 0.05). The estimated 5-year overall survival (OS) was 95.5% in the PCR group and 69.1% in the non-PCR group (p = 0.017). Overall, 11 patients (6.25%) presented with locoregional recurrence (LRR), one (1.6%) in the PCR group and 10 (8.8%) in the non-PCR group (p = 0.10). Conclusion: We observed higher PCR rates in triple-negative and HER2-positive molecular subtypes. DFS and OS were significantly better in patients who achieved PCR, regardless of clinicopathological features. We also observed lower rates of LRR in the population that reached PCR.

Analysis of time interval between breast cancer diagnosis and treatment / Análise do tempo decorrido entre diagnóstico e tratamento do câncer de mama

Introduction: Breast cancer has a good prognosis when treated early. However, the mortality rate in Brazil is still high. The time interval between radiological suspicion and diagnosis/treatment impacts the survival. Methods: This is a retrospective crosssectional study that assessed patients treated at a reference center, with abnormal breast imaging findings and subsequent confirmation of breast cancer, from January 2011 to June 2015. We reviewed variables related to the dates of the abnormal test result, first mastology appointment, biopsy, surgery, and the start of chemotherapy ­ when indicated. Time intervals were compared using the Friedman and Kruskal-Wallis tests with the software SPSS® 23.0. Results: We analyzed 65 patients. The median time between the abnormal test result and first mastology appointment was 35 days; between first mastology appointment and biopsy, 31 days; between biopsy and surgery, 85 days; and between surgery and chemotherapy, 137 days. The last two intervals showed significant differences (p<0.001). Discussion: Breast cancer patients had a significant delay until surgery and the start of chemotherapy. Early integration of the multidisciplinary team involved in this process and internal audits are necessary to optimize time intervals.

Introdução: O câncer de mama apresenta bom prognóstico quando tratado precocemente, entretanto, a mortalidade no Brasil continua elevada. O tempo entre suspeita radiológica e diagnóstico e tratamento tem impacto na sobrevida. Métodos: Foi realizado um estudo transversal e retrospectivo que avaliou pacientes atendidas em centro de referência com imagem mamária alterada e posterior confirmação de câncer de mama de janeiro de 2011 a junho e 2015. Foram revisadas variáveis relacionadas às datas do exame alterado, da primeira consulta, da biópsia, da cirurgia e do início da quimioterapia, quando indicada. Os intervalos de tempo foram comparados pelos testes Friedman e Kruskal-Wallis, pelo programa SPSS® 23.0. Resultados: Foram analisadas 65 pacientes. A mediana de tempo entre exame alterado e primeira consulta foi 35 dias, entre consulta na mastologia e biópsia foi 31 dias, entre biópsia e cirurgia foi 85 dias e entre cirurgia e quimioterapia foi 137 dias. Foram observadas diferenças significativas nos dois últimos intervalos (p<0,001). Discussão: As pacientes com câncer de mama apresentaram atraso significativo até a cirurgia e até o início da quimioterapia. Há necessidade da integração precoce da equipe multidisciplinar implicada nesse processo e auditorias internas a fim de otimizar os intervalos de tempo

ANDROGEN RECEPTOR EXPRESSION IN TRIPLE NEGATIVE BREAST CANCER AND ITS RELATIONSHIP TO PROGNOSTIC FACTORS / Expressão de receptor de androgênio em câncer de mama triplo negativo e sua relação com fatores prognósticos

Objective: Triple negative breast cancer (TNBC) is a subset of tumors with an aggressive intrinsic biology, resulting in poor prognosis. Androgen receptor (AR) is currently one of the most studied biomarkers in TNBC, playing a role in the genesis and development of breast cancer. Methods: In this cross-sectional study, we retrospectively reviewed the medical records of all patients with TNBC who received care from 2012 to 2014 at a single health center in southern Brazil. Histological material from breast tumors was analyzed by immunohistochemistry for AR expression and related to age, histological grade, tumor-infiltrating lymphocytes (TILs), and Ki-67. Results: Of 34 TNBC cases identified, 23 (67.6%) were AR negative and 11 (32.4%) were AR positive. The average age of the patients was 51.9 years (range: 30­82 years). Among positive cases, AR was weakly expressed in 6 and strongly expressed in 5 cases. Most patients (n=28; 82.0%) had poorly differentiated tumors. Mean Ki-67 expression was 65.0% in AR-negative and 43.6% in AR-positive cases (p<0.05). There was a significant association between age and AR expression (p<0.005), which was associated with mean age 70.8 years in the strongly AR-positive group and 42.3 years in the weakly AR-positive group. The mean percentage of TILs was 38.6% in AR-positive and 39.1% in AR-negative cases (p=0.391). Conclusion: There was no significant association between AR expression and histological grade or TILs. AR positivity in TNBC was associated with older age and tumors with lower Ki-67 expression, indicating two subgroups with distinct phenotypes in patients with TNBC.

Objetivo: O câncer de mama negativo triplo (triple negative breast cancer ­ TNBC) é um subtipo de tumores com biologia intrínseca agressiva, resultando em pior prognóstico. O receptor de andrógeno (androgen receptor ­ AR) é atualmente um dos biomarcadores mais estudados em TNBC, desempenhando papel na gênese e no desenvolvimento do câncer de mama. Métodos: Neste estudo transversal, revisamos retrospectivamente os registros médicos de todos os pacientes com TNBC que receberam atendimento de 2012 a 2014 em um único centro no sul do Brasil. O material histológico dos tumores de mama foi analisado por imuno-histoquímica para a expressão de AR e relacionado a idade, grau histológico, linfócitos infiltrantes de tumores (TILs) e Ki-67. Resultados: Dos 34 casos identificados de TNBC, 23 (67,6%) eram AR negativos e 11 (32,4%), AR positivos. A idade média foi de 51,9 anos (30­82 anos). Entre os casos positivos, AR foi fracamente expresso em 6 e fortemente expresso em 5 casos. A maioria dos pacientes (n=28, 82,0%) apresentou tumores pouco diferenciados. A expressão média de Ki-67 foi de 65,0% em AR-negativo e 43,6% em AR-positivo (p<0,05). Houve associação significativa entre a idade e a expressão de AR (p<0,005), associada à idade média de 70,8 anos no grupo com AR fortemente positivo e de 42,3 anos no grupo com AR fracamente positivo. A porcentagem média de TILs foi de 38,6% em AR-positivo e de 39,1% em AR-negativo (p=0,391). Não houve associação significativa entre expressão AR e grau histológico ou TILs. Conclusão: A positividade de AR em TNBC foi associada com idade mais avançada e tumores com menor expressão de Ki-67, indicando dois subgrupos com fenótipos distintos em pacientes com TNBC.