RESUMO
BACKGROUND: Although the role of viral agents, such as human papillomavirus (e.g. HPV16, HPV18) in colorectal cancer (CRC) has been previously investigated, results remain inconclusive. METHODS: To further evaluate the involvement of oncogenic HPV types in CRC, 40 frozen neoplastic and 40 adjacent colonic tissues collected from Italian patients were analyzed by Luminex-based assays that detect a broad spectrum of HPV types, i.e. Alpha (n = 21), Beta (n = 46) and Gamma HPVs (n = 52). In addition, 125 frozen CRC samples and 70 surrounding mucosal tissues were collected from Czech patients and analyzed by broad spectrum PCR protocols: (i) FAP59/64, (ii) FAPM1 and (iii) CUT combined with Next Generation Sequencing (NGS). RESULTS: Using Luminex-basedassays, DNA from HPV16 was detected in 5% (2/40) CRC tissues from Italian patients. One HPV16 DNA-positive CRC case was subsequently confirmed positive for E6*I mRNA. Cutaneous beta HPV types were detected in 10% (4/40) adjacent tissues only, namely HPV111 (n = 3) and HPV120 (n = 1), while gamma HPV168 (n = 1) and HPV199 (n = 1) types were detected in adjacent and in tumor tissues, respectively. The NGS analysis of the CRC Czech samples identified HPV sequences from mucosal alpha-3 (HPV89), alpha-7 (HPV18, 39, 68 and 70) and alpha-10 species (HPV11), as well as cutaneous beta-1 (HPV20, 24, 93, 98, 105,124) beta-2 (HPV23), beta-3 (HPV49) and gamma-1 species (HPV205). CONCLUSIONS: Our findings indicate that HPV types belonging to the mucosal alpha, and the 'cutaneous' beta and gamma genera can be detected in the colonic mucosal samples with a low prevalence rate and a low number of HPV reads by Luminex and NGS, respectively. However, additional studies are required to corroborate these findings.
RESUMO
Introduction: Recently, accurate machine learning and deep learning approaches have been dedicated to the investigation of breast cancer invasive disease events (IDEs), such as recurrence, contralateral and second cancers. However, such approaches are poorly interpretable. Methods: Thus, we designed an Explainable Artificial Intelligence (XAI) framework to investigate IDEs within a cohort of 486 breast cancer patients enrolled at IRCCS Istituto Tumori "Giovanni Paolo II" in Bari, Italy. Using Shapley values, we determined the IDE driving features according to two periods, often adopted in clinical practice, of 5 and 10 years from the first tumor diagnosis. Results: Age, tumor diameter, surgery type, and multiplicity are predominant within the 5-year frame, while therapy-related features, including hormone, chemotherapy schemes and lymphovascular invasion, dominate the 10-year IDE prediction. Estrogen Receptor (ER), proliferation marker Ki67 and metastatic lymph nodes affect both frames. Discussion: Thus, our framework aims at shortening the distance between AI and clinical practice.
RESUMO
During Covid-19 pandemic most hospitals have restricted in-person delivery of non-essential healthcare services, including genetic testing delivery, to slow the spread of the virus. Our Onco-Genetic Service also faced this challenging period and had to re-organize its clinical practice with the use of tele-health. Aim of the present paper is to understand whether and how Covid-19-related changes in medical practice influenced patients' satisfaction about the health service provided. 125 BRCA1/2 non carriers (109/125, 87.2% female and 16/125, 12.8% male) in Istituto Tumori "Giovanni Paolo II" of Bari were enrolled. All participants were asked to choose whether they prefer in-person or remote post-test counselling session. Basing on patients' choice, two groups of subjects were composed. One week after the post-test counselling session, participants were phone called and asked to complete: a socio-demographic form, a brief structured interview about their Covid-19 related worries and their satisfaction with the health service provided, Hospital Anxiety and Depression Scale and Fear of Covid-19 scale. Qualitative information about patients' choice were also collected. No significant difference about patients' satisfaction with the health service provided emerged between groups. Patients who preferred remote post-test counselling had higher anxiety, worries and fear-of Covid-19 than the others. All remote-counselling subjects preferred tele-genetics because of Covid-19 security, would choose it again and would recommend it to others. Cancer tele-genetics offers good guarantees of comfort and efficacy, but patients' choices are related to personal and psychological variables. The use of tele-genetics has to be a patient's choice.
Assuntos
Proteína BRCA1 , Proteína BRCA2 , COVID-19 , Testes Genéticos , Feminino , Humanos , Masculino , Ansiedade/psicologia , Proteína BRCA1/genética , COVID-19/epidemiologia , Pandemias , Satisfação do Paciente , Proteína BRCA2/genética , Consulta Remota , Preferência do PacienteRESUMO
BACKGROUND AND AIM OF THE WORK: BRCA1/2 are tumour-suppressor genes involved in DNA homologous recombination and ovarian cancer development. The study evaluated the risk of tumor cancer in women presenting the BRCA mutations. METHODS: Risk-reducing surgery (RRS) was performed in 100 patients carrying BRCA1 (aged between 30-73 years, median age was 51 years) and BRCA 2 mutation (aged between 36-70 years, median age was 53 years). Fifty-eight patients had previous history of breast cancer. RESULTS: Between the 100 patients, 82 women underwent risk-reducing salpingo-oophorectomy (RRSO) through a laparoscopic minimally invasive approach, 7 (7 %) underwent laparoscopic RRSO and contextual hysterectomy, 1 woman (1 %) underwent RRSO through a laparotomic approach and 10 women (10 %) laparotomic RRSO and hysterectomy. During 5 (5 %) laparoscopic RRSO, prophylactic bilateral mastectomy was also performed. Early and late complication occurred in 3 patients (3 %). Two patients (2 %) were found to have occult Serous Tubal Intraepithelial Carcinoma (STIC) and three patients (3 %) occult cancer. CONCLUSIONS: RRSO is safe and feasible in BRCA mutation carriers. The procedure is effective for genetic prevention of ovarian cancer.
Assuntos
Neoplasias da Mama , Neoplasias Ovarianas , Adulto , Idoso , Neoplasias da Mama/genética , Neoplasias da Mama/prevenção & controle , Neoplasias da Mama/cirurgia , Feminino , Genes BRCA1 , Genes BRCA2 , Humanos , Mastectomia , Pessoa de Meia-Idade , Mutação , Neoplasias Ovarianas/genética , Neoplasias Ovarianas/prevenção & controle , Neoplasias Ovarianas/cirurgia , Ovariectomia , Estudos Prospectivos , Salpingo-OoforectomiaRESUMO
Although research biobanks are among the most promising tools to fight disease and improve public health, there are a range of risks biobanks may face that mainly need to be assessed in an attempt to be relieved. We conducted a strategic insurance review of an institutional cancer biobank with the aim of both identifying the insurable risks of our own Biobank and gathering useful evidence of primary exposure to insurable risks. In this practical scenario, risks have been outlined and categorized into inherent and residual risks, along with their possible impact on biobank maintenance. Results at the Biobank of the Cancer Institute of Bari showed evidence of potentially significant and intrinsic risk due to highly relevant threats, along with already implemented improvements that significantly reduce risks to a range of relative acceptability.
RESUMO
To date, some artificial intelligence (AI) methods have exploited Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) to identify finer tumor properties as potential earlier indicators of pathological Complete Response (pCR) in breast cancer patients undergoing neoadjuvant chemotherapy (NAC). However, they work either for sagittal or axial MRI protocols. More flexible AI tools, to be used easily in clinical practice across various institutions in accordance with its own imaging acquisition protocol, are required. Here, we addressed this topic by developing an AI method based on deep learning in giving an early prediction of pCR at various DCE-MRI protocols (axial and sagittal). Sagittal DCE-MRIs refer to 151 patients (42 pCR; 109 non-pCR) from the public I-SPY1 TRIAL database (DB); axial DCE-MRIs are related to 74 patients (22 pCR; 52 non-pCR) from a private DB provided by Istituto Tumori "Giovanni Paolo II" in Bari (Italy). By merging the features extracted from baseline MRIs with some pre-treatment clinical variables, accuracies of 84.4% and 77.3% and AUC values of 80.3% and 78.0% were achieved on the independent tests related to the public DB and the private DB, respectively. Overall, the presented method has shown to be robust regardless of the specific MRI protocol.
RESUMO
Cancer treatment related infertility (CTRI) affects more than one third of young women undergoing anti-cancer protocols, inducing a premature exhaustion of the ovarian reserve. In addition to ovarian suppression by GnRHa, oocyte and cortex cryopreservation has gained interest in patients with estrogen-sensitive tumors for whom the hormonal burst to prompt the multiple follicular growth could provide a further pro-life tumor pulsing. On the other hand, cortex reimplantation implies a few drawbacks due to the unknown consistency of the follicles to be reimplanted or the risk of reintroducing malignant cells. The capability of ovarian stem cells (OCSs) from fresh ovarian cortex fragments to differentiate in vitro to mature oocytes provides a tool to overcome these drawbacks. In fact, since ovarian cortex sampling and cryopreservation is practicable before gonadotoxic treatments, the recruitment of OSCs from defrosted fragments could provide a novel opportunity to verify their suitability to be expanded in vitro as oocyte like cells (OLCs). Here, we describe in very preliminary experiments the consistency of an OSC population from a single cryopreserved ovarian cortex after thawing as well as both their viability and their suitability to be further explored in their property to differentiate in OLCs, thus reinforcing interest in stemness studies in the treatment of female CTRI.
RESUMO
ASTRACT: The usefulness of this review is to highlight how a fertility preservation (FP) approach is currently feasible for patients diagnosed with uterine cervical cancer. To this regard, a fertility sparing surgery has just overcome its traditional limits, gained acceptance within the major gynecologic oncology societies thanks to the ability to identify the "ideal" candidates to this conservative treatment. On the other hand, the use of other FPs for oocyte and ovarian cortex cryopreservation is still extremely debated. In fact, the existing risk of tumor spreading during oocyte retrieval necessary for oocyte cryostorage for patients' candidates for neo-adjuvant therapy, as well as the potential hazard of cancer cell dissemination after ovarian tissue replacement in cases of non-squamous type cervical carcinomas should not be underestimated. Therefore, in consideration of the encountered limitations and the need to ensure adequate reproductive health for young uterine cervical cancer survivors, translational research regarding the FP has progressively collected innovative insights into the employment of stemness technology. In this context, the property of ovarian stem cells obtained from the ovarian cortex to generate functional oocytes in women could represent a promising therapeutic alternative to the current procedures for a novel and safer FP approach in cancer survivors.
Assuntos
Carcinoma , Preservação da Fertilidade , Neoplasias do Colo do Útero , Criopreservação/métodos , Feminino , Preservação da Fertilidade/métodos , Humanos , Recuperação de Oócitos , Oócitos , Neoplasias do Colo do Útero/cirurgiaRESUMO
The female carriers of BRCA1/2 pathogenic variants (mutations) face a high lifetime risk of developing breast and/or ovarian cancer. However, the risk may differ depending on various genetic and non-genetic elements, including metabolic and hormonal factors. We previously showed that a 6-month Mediterranean dietary intervention trial reduced body weight and the levels of insulin-like growth factor I and other metabolic factors in BRCA mutation carriers. We also found that higher baseline levels of glucose and insulin were significantly associated with BRCA loss-of-function (LOF) variants. In this study, we evaluated whether the BRCA mutation type influences in a different way the metabolic and hormonal response to the dietary intervention in 366 female carriers. The LOF variant carriers randomized in the intervention group (IG) showed significantly higher changes in most considered parameters compared to the control group (CG). The nonsynonymous variant carriers in the IG showed similar changes, but none of them were statistically significant. Performing the "delta" analysis of differences (intention-to-treat analysis), we observed that in LOF variant carriers, the reduction of insulin levels was significantly more pronounced that in nonsynonymous variant carriers. These findings suggest that the changes in insulin levels might be modulated by a different response to the dietary intervention mediated by BRCA LOF variants.
RESUMO
BACKGROUND: The immunotherapy with immune checkpoints inhibitors (ICI) has changed the life expectancy in metastatic melanoma (MM) patients. Nevertheless, several patients do not respond hence, the identification and validation of novel biomarkers of response to ICI is of crucial importance. Circulating extracellular vesicles (EVs) such as PD-L1+ EV mediate resistance to anti-PD1, instead the role of PD1+ EV is not fully understood. METHODS: We isolated the circulating EVs from the plasma of an observational cohort study of 71 metastatic melanoma patients and correlated the amount of PD-L1+ EVs and PD1+ EVs with the response to ICI. The analysis was performed according to the origin of EVs from the tumor and the immune cells. Subsequently, we analysed the data in a validation cohort of 22 MM patients to assess the reliability of identified EV-based biomarkers. Additionally we assessed the involvement of PD1+ EVs in the seizure of nivolumab and in the perturbation of immune cells-mediated killing of melanoma spheroids. RESULTS: The level of PD-L1+ EVs released from melanoma and CD8+ T cells and that of PD1+ EVs irrespective of the cellular origin were higher in non-responders. The Kaplan-Meier curves indicated that higher levels of PD1+ EVs were significantly correlated with poorer progression-free survival (PFS) and overall survival (OS). Significant correlations were found for PD-L1+ EVs only when released from melanoma and T cells. The multivariate analysis showed that high level of PD1+ EVs, from T cells and B cells, and high level of PD-L1+ EVs from melanoma cells, are independent biomarkers of response. The reliability of PD-L1+ EVs from melanoma and PD1+ EVs from T cells in predicting PFS was confirmed in the validation cohort through the univariate Cox-hazard regression analysis. Moreover we discovered that the circulating EVs captured nivolumab and reduced the T cells trafficking and tumor spheroids killing. CONCLUSION: Our study identified circulating PD1+ EVs as driver of resistance to anti-PD1, and highlighted that the analysis of single EV population by liquid biopsy is a promising tool to stratify MM patients for immunotherapy.
Assuntos
Antígeno B7-H1/metabolismo , Biomarcadores Tumorais , Resistencia a Medicamentos Antineoplásicos , Vesículas Extracelulares/metabolismo , Melanoma/metabolismo , Receptor de Morte Celular Programada 1/metabolismo , Antígeno B7-H1/genética , Diagnóstico por Imagem , Resistencia a Medicamentos Antineoplásicos/genética , Feminino , Humanos , Inibidores de Checkpoint Imunológico/farmacologia , Inibidores de Checkpoint Imunológico/uso terapêutico , Imunofenotipagem , Masculino , Melanoma/diagnóstico , Melanoma/tratamento farmacológico , Melanoma/etiologia , Metástase Neoplásica , Estadiamento de Neoplasias , Receptor de Morte Celular Programada 1/genética , Modelos de Riscos Proporcionais , Reprodutibilidade dos TestesRESUMO
BRCA1/2-associated hereditary breast and ovarian cancer is the most common form of hereditary breast and ovarian cancer and occurs in all ethnicities and racial populations. Different BRCA1/BRCA2 pathogenic variants (PVs) have been reported with a wide variety among populations. In this study, we retrospectively analyzed prevalence and geographic distribution of pathogenic germline BRCA1/2 variants in families from Apulia in southern Italy and evaluated the genotype-phenotype correlations. Data were collected from Oncogenetic Services present in Apulian hospitals and a shared database was built containing Apulian native probands (n = 2026) that had undergone genetic testing from 2004 to 2019. PVs were detected in 499 of 2026 (24.6%) probands and 68.5% of them (342 of 499) were in the BRCA1 gene. We found 65 different PVs in BRCA1 and 46 in BRCA2. There were 10 most recurrent PVs and their geographical distribution appears to be significantly specific for each province. We have assumed that these PVs are related to the historical and geopolitical changes that occurred in Apulia over time and/or to a "founder effect". Broader knowledge of BRCA1/2 prevalence and recurring PVs in specific geographic areas could help establish more flexible genetic testing strategies that may enhance our ability to detect high-risk subjects.
RESUMO
The coronavirus disease-2019 (COVID-19) pandemic has caused an enormous loss of lives. Various clinical trials of vaccines and drugs are being conducted worldwide; nevertheless, as of today, no effective drug exists for COVID-19. The identification of key genes and pathways in this disease may lead to finding potential drug targets and biomarkers. Here, we applied weighted gene co-expression network analysis and LIME as an explainable artificial intelligence algorithm to comprehensively characterize transcriptional changes in bronchial epithelium cells (primary human lung epithelium (NHBE) and transformed lung alveolar (A549) cells) during severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Our study detected a network that significantly correlated to the pathogenicity of COVID-19 infection based on identified hub genes in each cell line separately. The novel hub gene signature that was detected in our study, including PGLYRP4 and HEPHL1, may shed light on the pathogenesis of COVID-19, holding promise for future prognostic and therapeutic approaches. The enrichment analysis of hub genes showed that the most relevant biological process and KEGG pathways were the type I interferon signaling pathway, IL-17 signaling pathway, cytokine-mediated signaling pathway, and defense response to virus categories, all of which play significant roles in restricting viral infection. Moreover, according to the drug-target network, we identified 17 novel FDA-approved candidate drugs, which could potentially be used to treat COVID-19 patients through the regulation of four hub genes of the co-expression network. In conclusion, the aforementioned hub genes might play potential roles in translational medicine and might become promising therapeutic targets. Further in vitro and in vivo experimental studies are needed to evaluate the role of these hub genes in COVID-19.
RESUMO
Multiple myeloma (MM) is a hematological malignancy characterised by proliferation of clonal plasma cells (PCs) within the bone marrow (BM). Myelomagenesis is a multi-step process which goes from an asymptomatic phase, defined as monoclonal gammopathy of undetermined significance (MGUS), to a smouldering myeloma (SMM) stage, to a final active MM disease, characterised by hypercalcemia, renal failure, bone lesions anemia, and higher risk of infections. Overall, microRNAs (miRNAs) have shown to significantly impact on MM tumorigenesis, as a result of miRNA-dependent modulation of genes involved in pathways known to be crucial for MM pathogenesis and disease progression. We aim to revise the literature related to the role of miRNAs as potential diagnostic and prognostic biomarkers, thus highlighting their key role as novel players within the field of MM and related premalignant conditions.
RESUMO
In the second reconstructive phase of the breast after mastectomy, lipofilling is often necessary. Currently, lipofilling occurs immediately after autologous adipose tissue harvesting procedure, but most of the patients, usually, require multiple sessions to obtain a satisfactory result. Therefore, the need of repeated surgical harvesting outputs implies high risk of patients' morbidity and discomfort as well as increasing medical time and costs. The aim of our pilot study was to find out a feasible method to cryopreserve adipose tissue, in order to avoid reiterated liposuctions. Lipoaspirates samples have been harvested from 10 women and preserved by three methods: (1) the first one, using 10% Me2SO and 20% human albumin from human plasma as cryoprotective agents; (2) the second one, adding 5% Me2SO as cryoprotective agent; 3) the last one, without any cryoprotective agent. Fresh and cryopreserved fat samples, obtained through the aforementioned processes, have been analyzed ex vivo. The efficiency of the cryopreservation methods used was determined by adipocyte viability and the expression of adipocytes surface markers. Lipoaspirates stored at -196 °C for 3 months, after thawing, retained comparable adipocyte viability and histology to fresh tissue and no significant differences were found between the three methods used. Although the current results, differences between the methodologies in terms of viability may not become evident until breast lipofilling using frozen-thawed cryopreserved tissue.
Assuntos
Neoplasias da Mama , Crioprotetores , Tecido Adiposo , Criopreservação/métodos , Crioprotetores/farmacologia , Dimetil Sulfóxido , Feminino , Humanos , Mastectomia , Projetos PilotoRESUMO
An important group of breast cancers is those associated with inherited susceptibility. In women, several predisposing mutations in genes involved in DNA repair have been discovered. Women with a germline pathogenic variant in BRCA1 have a lifetime cancer risk of 70%. As part of a larger prospective study on heavy metals, our aim was to investigate if blood arsenic levels are associated with breast cancer risk among women with inherited BRCA1 mutations. A total of 1084 participants with pathogenic variants in BRCA1 were enrolled in this study. Subjects were followed from 2011 to 2020 (mean follow-up time: 3.75 years). During that time, 90 cancers were diagnosed, including 67 breast and 10 ovarian cancers. The group was stratified into two categories (lower and higher blood As levels), divided at the median (<0.85 µg/L and ≥0.85 µg/L) As level among all unaffected participants. Cox proportional hazards models were used to model the association between As levels and cancer incidence. A high blood As level (≥0.85 µg/L) was associated with a significantly increased risk of developing breast cancer (HR = 2.05; 95%CI: 1.18-3.56; p = 0.01) and of any cancer (HR = 1.73; 95%CI: 1.09-2.74; p = 0.02). These findings suggest a possible role of environmental arsenic in the development of cancers among women with germline pathogenic variants in BRCA1.
RESUMO
The term "biobanking" is often misapplied to any collection of human biological materials (biospecimens) regardless of requirements related to ethical and legal issues or the standardization of different processes involved in tissue collection. A proper definition of biobanks is large collections of biospecimens linked to relevant personal and health information (health records, family history, lifestyle, genetic information) that are held predominantly for use in health and medical research. In addition, the International Organization for Standardization, in illustrating the requirements for biobanking (ISO 20387:2018), stresses the concept of biobanks being legal entities driving the process of acquisition and storage together with some or all of the activities related to collection, preparation, preservation, testing, analysing and distributing defined biological material as well as related information and data. In this review article, we aim to discuss the basic principles of biobanking, spanning from definitions to classification systems, standardization processes and documents, sustainability and ethical and legal requirements. We also deal with emerging specimens that are currently being generated and shaping the so-called next-generation biobanking, and we provide pragmatic examples of cancer-associated biobanking by discussing the process behind the construction of a biobank and the infrastructures supporting the implementation of biobanking in scientific research.
Assuntos
Bancos de Espécimes Biológicos , Pesquisa Biomédica , Medicina de Precisão , Manejo de Espécimes , Acreditação , Bancos de Espécimes Biológicos/classificação , Bancos de Espécimes Biológicos/ética , Bancos de Espécimes Biológicos/legislação & jurisprudência , Bancos de Espécimes Biológicos/normas , Pesquisa Biomédica/classificação , Pesquisa Biomédica/ética , Pesquisa Biomédica/legislação & jurisprudência , Pesquisa Biomédica/normas , Guias como Assunto , Humanos , Formulação de Políticas , Medicina de Precisão/classificação , Medicina de Precisão/ética , Medicina de Precisão/normas , Manejo de Espécimes/classificação , Manejo de Espécimes/ética , Manejo de Espécimes/normas , Participação dos Interessados , Terminologia como AssuntoRESUMO
The dynamic contrast-enhanced MR imaging plays a crucial role in evaluating the effectiveness of neoadjuvant chemotherapy (NAC) even since its early stage through the prediction of the final pathological complete response (pCR). In this study, we proposed a transfer learning approach to predict if a patient achieved pCR (pCR) or did not (non-pCR) by exploiting, separately or in combination, pre-treatment and early-treatment exams from I-SPY1 TRIAL public database. First, low-level features, i.e., related to local structure of the image, were automatically extracted by a pre-trained convolutional neural network (CNN) overcoming manual feature extraction. Next, an optimal set of most stable features was detected and then used to design an SVM classifier. A first subset of patients, called fine-tuning dataset (30 pCR; 78 non-pCR), was used to perform the optimal choice of features. A second subset not involved in the feature selection process was employed as an independent test (7 pCR; 19 non-pCR) to validate the model. By combining the optimal features extracted from both pre-treatment and early-treatment exams with some clinical features, i.e., ER, PgR, HER2 and molecular subtype, an accuracy of 91.4% and 92.3%, and an AUC value of 0.93 and 0.90, were returned on the fine-tuning dataset and the independent test, respectively. Overall, the low-level CNN features have an important role in the early evaluation of the NAC efficacy by predicting pCR. The proposed model represents a first effort towards the development of a clinical support tool for an early prediction of pCR to NAC.
Assuntos
Neoplasias da Mama/diagnóstico , Neoplasias da Mama/tratamento farmacológico , Mama/diagnóstico por imagem , Imageamento por Ressonância Magnética , Adulto , Mama/efeitos dos fármacos , Mama/patologia , Neoplasias da Mama/diagnóstico por imagem , Neoplasias da Mama/genética , Feminino , Humanos , Aprendizado de Máquina , Pessoa de Meia-Idade , Estadiamento de Neoplasias , Redes Neurais de Computação , Radiografia , Receptor ErbB-2/genética , Receptores de Estrogênio/genética , Receptores de Progesterona/genética , Resultado do TratamentoRESUMO
Cancer treatment planning benefits from an accurate early prediction of the treatment efficacy. The goal of this study is to give an early prediction of three-year Breast Cancer Recurrence (BCR) for patients who underwent neoadjuvant chemotherapy. We addressed the task from a new perspective based on transfer learning applied to pre-treatment and early-treatment DCE-MRI scans. Firstly, low-level features were automatically extracted from MR images using a pre-trained Convolutional Neural Network (CNN) architecture without human intervention. Subsequently, the prediction model was built with an optimal subset of CNN features and evaluated on two sets of patients from I-SPY1 TRIAL and BREAST-MRI-NACT-Pilot public databases: a fine-tuning dataset (70 not recurrent and 26 recurrent cases), which was primarily used to find the optimal subset of CNN features, and an independent test (45 not recurrent and 17 recurrent cases), whose patients had not been involved in the feature selection process. The best results were achieved when the optimal CNN features were augmented by four clinical variables (age, ER, PgR, HER2+), reaching an accuracy of 91.7% and 85.2%, a sensitivity of 80.8% and 84.6%, a specificity of 95.7% and 85.4%, and an AUC value of 0.93 and 0.83 on the fine-tuning dataset and the independent test, respectively. Finally, the CNN features extracted from pre-treatment and early-treatment exams were revealed to be strong predictors of BCR.
RESUMO
Health-care workers (HCW) are at high risk for SARS-CoV-2 infection and, if asymptomatic, for transmitting the virus to fragile cancer patients. We monitored all asymptomatic HCWs of a cancer institute (94% of all employees agreed to enter the study) with the rapid serological test, VivaDiagTM, identifying SARS-CoV-2 associated-IgM/IgG. The tests were performed at time 0 (n = 606) and after 14 days (n = 393). Overall, the VivaDiagTM results of nine HCWs (1.5%) were positive, with one confirmed to be SARS-CoV-2-positive after oropharyngeal swab testing by RT-PCR. At time 0, all nine cases showed IgM expression while IgG was detected in only one. After 14 days, IgM persisted in all the cases, while IgG became evident in four. A chemiluminescence immunoassay (CLIA) confirmed IgM positivity in 5/13 VivaDiagTM positive cases and IgG positivity in 4/5 VivaDiagTM positive cases. Our study suggests that the VivaDiagTM test can be of help in identifying SARS-CoV-2 infected people in cohorts of subjects with a high prevalence.