Construction and validation of a deep learning prognostic model based on digital pathology images of stage III colorectal cancer.

BACKGROUND: TNM staging is the main reference standard for prognostic prediction of colorectal cancer (CRC), but the prognosis heterogeneity of patients with the same stage is still large. This study aimed to classify the tumor microenvironment of patients with stage III CRC and quantify the classified tumor tissues based on deep learning to explore the prognostic value of the developed tumor risk signature (TRS). METHODS: A tissue classification model was developed to identify nine tissues (adipose, background, debris, lymphocytes, mucus, smooth muscle, normal mucosa, stroma, and tumor) in whole-slide images (WSIs) of stage III CRC patients. This model was used to extract tumor tissues from WSIs of 265 stage III CRC patients from The Cancer Genome Atlas and 70 stage III CRC patients from the Sixth Affiliated Hospital of Sun Yat-sen University. We used three different deep learning models for tumor feature extraction and applied a Cox model to establish the TRS. Survival analysis was conducted to explore the prognostic performance of TRS. RESULTS: The tissue classification model achieved 94.4 % accuracy in identifying nine tissue types. The TRS showed a Harrell's concordance index of 0.736, 0.716, and 0.711 in the internal training, internal validation, and external validation sets. Survival analysis showed that TRS had significant predictive ability (hazard ratio: 3.632, p = 0.03) for prognostic prediction. CONCLUSION: The TRS is an independent and significant prognostic factor for PFS of stage III CRC patients and it contributes to risk stratification of patients with different clinical stages.

Radioport: a radiomics-reporting network for interpretable deep learning in BI-RADS classification of mammographic calcification.

Objective.Generally, due to a lack of explainability, radiomics based on deep learning has been perceived as a black-box solution for radiologists. Automatic generation of diagnostic reports is a semantic approach to enhance the explanation of deep learning radiomics (DLR).Approach.In this paper, we propose a novel model called radiomics-reporting network (Radioport), which incorporates text attention. This model aims to improve the interpretability of DLR in mammographic calcification diagnosis. Firstly, it employs convolutional neural networks to extract visual features as radiomics for multi-category classification based on breast imaging reporting and data system. Then, it builds a mapping between these visual features and textual features to generate diagnostic reports, incorporating an attention module for improved clarity.Main results.To demonstrate the effectiveness of our proposed model, we conducted experiments on a breast calcification dataset comprising mammograms and diagnostic reports. The results demonstrate that our model can: (i) semantically enhance the interpretability of DLR; and, (ii) improve the readability of generated medical reports.Significance.Our interpretable textual model can explicitly simulate the mammographic calcification diagnosis process.

Near-Infrared-II Nanoparticles for Vascular Normalization Combined with Immune Checkpoint Blockade via Photodynamic Immunotherapy Inhibit Uveal Melanoma Growth and Metastasis.

Photodynamic therapy (PDT) has been widely employed in tumor treatment due to its effectiveness. However, the tumor hypoxic microenvironment which is caused by abnormal vasculature severely limits the efficacy of PDT. Furthermore, the abnormal vasculature has been implicated in the failure of immunotherapy. In this study, a novel nanoparticle denoted as Combo-NP is introduced, composed of a biodegradable NIR II fluorescent pseudo-conjugate polymer featuring disulfide bonds within its main chain, designated as TPA-BD, and the vascular inhibitor Lenvatinib. Combo-NP exhibits dual functionality by not only inducing cytotoxic reactive oxygen species (ROS) to directly eliminate tumor cells but also eliciting immunogenic cell death (ICD). This ICD response, in turn, initiates a robust cascade of immune reactions, thereby augmenting the generation of cytotoxic T lymphocytes (CTLs). In addition, Combo-NP addresses the issue of tumor hypoxia by normalizing the tumor vasculature. This normalization process enhances the efficacy of PDT while concurrently fostering increased CTLs infiltration within the tumor microenvironment. These synergistic effects synergize to potentiate the photodynamic-immunotherapeutic properties of the nanoparticles. Furthermore, when combined with anti-programmed death-ligand 1 (PD-L1), they showcase notable inhibitory effects on tumor metastasis. The findings in this study introduce an innovative nanomedicine strategy aimed at triggering systemic anti-tumor immune responses for the treatment of Uveal melanoma.

Attention mechanism based multi-sequence MRI fusion improves prediction of response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer.

BACKGROUND: Accurate prediction of response to neoadjuvant chemoradiotherapy (nCRT) is very important for treatment plan decision in locally advanced rectal cancer (LARC). The aim of this study was to investigate whether self-attention mechanism based multi-sequence fusion strategy applied to multiparametric magnetic resonance imaging (MRI) based deep learning or hand-crafted radiomics model construction can improve prediction of response to nCRT in LARC. METHODS: This retrospective analysis enrolled 422 consecutive patients with LARC who received nCRT before surgery at two hospitals. All patients underwent multiparametric MRI scans with three imaging sequences. Tumor regression grade (TRG) was used to assess the response of nCRT based on the resected specimen. Patients were separated into 2 groups: poor responders (TRG 2, 3) versus good responders (TRG 0, 1). A self-attention mechanism, namely channel attention, was applied to fuse the three sequence information for deep learning and radiomics models construction. For comparison, other two models without channel attention were also constructed. All models were developed in the same hospital and validated in the other hospital. RESULTS: The deep learning model with channel attention mechanism achieved area under the curves (AUCs) of 0.898 in the internal validation cohort and 0.873 in the external validation cohort, which was the best performed model in all cohorts. More importantly, both the deep learning and radiomics model that applied channel attention mechanism performed better than those without channel attention mechanism. CONCLUSIONS: The self-attention mechanism based multi-sequence fusion strategy can improve prediction of response to nCRT in LARC.

Comparison of clinical outcomes between cystotome-assisted prechop phacoemulsification surgery and femtosecond laser-assisted cataract surgery for hard nucleus cataracts.

BACKGROUND/OBJECTIVES: To compare the safety and efficacy of cystotome-assisted prechop phacoemulsification surgery (CAPPS) and femtosecond laser-assisted cataract surgery (FLACS) in patients with hard nucleus cataract. SUBJECTS/METHODS: Ninety-six eyes of 64 patients with grade IV hard nucleus cataract were assigned to 1 of the 2 groups (49 CAPPS and 47 FLACS). Follow-up visits were performed at 1 day, 1 week, 1 month, 3 months, 6 months, and 1 year, and the outcome measures comprised ultrasound power, effective phacoemulsification time (EPT), corrected distance visual acuity (CDVA), endothelial cell density (ECD), corneal endothelium cell loss rate (ECL), central corneal thickness (CCT), and intraoperative and postoperative complications. RESULTS: The ultrasound power and EPT were lower in the CAPPS group (p = 0.03 and <0.0001, respectively). Patients in both groups gained better CDVA postoperatively. The ECD value decreased at each follow-up visit and did not return to the preoperative level; FLACS resulted in greater endothelial cell loss compared to CAPPS. CCT increased immediately after the surgery and decreased thereafter. The mean CCT value returned to the preoperative level 3 months postoperatively in the CAPPS group, while in the FLACS group, CCT value took 6 months to return to the preoperative level. Miosis was more likely to occur in the FLACS group. CONCLUSIONS: Due to its efficacy and cost-effectiveness, CAPPS is worth promoting and applying to clinical work in the future.

Integrated analysis to reveal potential therapeutic targets and prognostic biomarkers of skin cutaneous melanoma.

Skin cutaneous melanoma (SKCM) is a malignant tumor with high mortality rate in human, and its occurrence and development are jointly regulated by genes and the environment. However, the specific pathogenesis of SKCM is not completely understood. In recent years, an increasing number of studies have reported the important role of competing endogenous RNA (ceRNA) regulatory networks in various tumors; however, the complexity and specific biological effects of the ceRNA regulatory network of SKCM remain unclear. In the present study, we obtained a ceRNA regulatory network of long non-coding RNAs, microRNAs, and mRNAs related to the phosphatase and tensin homolog (PTEN) in SKCM and identified the potential diagnostic and prognostic markers related to SKCM. We extracted the above three types of RNA involved in SKCM from The Cancer Genome Atlas database. Through bioinformatics analysis, the OIP5-AS1-hsa-miR-186-5p/hsa-miR-616-3p/hsa-miR-135a-5p/hsa-miR-23b-3p/hsa-miR-374b-5p-PTPRC/IL7R/CD69 and MALAT1-hsa-miR-135a-5p/hsa-miR-23b-3p/hsa-miR-374b-5p-IL7R/CD69 ceRNA networks were found to be related to the prognosis of SKCM. Finally, we determined the OIP5-AS1-PTPRC/IL7R/CD69 and MALAT1-IL7R/CD69 axes in ceRNA as a clinical prognostic model using correlation and Cox regression analyses. Additionally, we explored the possible role of these two axes in affecting gene expression and immune microenvironment changes and the occurrence and development of SKCM through methylation and immune infiltration analyses. In summary, the ceRNA-based OIP5-AS1-PTPRC/IL7R/CD69 and MALAT1-IL7R/CD69 axes may be a novel and important approach for the diagnosis and prognosis of SKCM.

Factors Influencing the Acceptance of Endoscopic Thyroidectomy Via Oral Vestibular Approach (ETOVA) by Thyroid Surgery Candidates.

OBJECTIVE: To investigate the factors affecting the acceptance of endoscopic thyroidectomy via the oral vestibular approach (ETOVA) in Chinese patients before thyroid surgery. METHODS: The enrolled patients were asked to answer a questionnaire postoperatively about their demographics, medical insurance coverage, sources of information, reasons for selection, and safety. The relationship between the collected data and the acceptance of ETOVA was analyzed. RESULTS: Two hundred patients (40 males, 20%; 160 females, 80%) answered the questionnaire. One hundred sixty-two of them (81%) accepted ETOVA. Univariate analysis showed that the patients' age, cosmetic effect, safety, results perception, and recommendations from family, friends, doctors, and nurses are correlated with the acceptance of ETOVA. Multivariate analysis showed that patients' age (OR=0.966, P =0.015), cosmetic effect (OR=12.620, P =0.000), safety (OR=0.295, P =0.016), minimal invasion (OR=4.877, P =0.001), and doctors/nurses' advance (OR=4.485, P =0.017) are statistically significant and were positively correlated with the acceptance of ETOVA. Education level, medical insurance coverage, family support, past surgical history, and operative-related symptoms were not statistically significant ( P >0.05). CONCLUSION: Among thyroid surgery candidates in Southwest China, younger patients with cosmetic requirements and minimally invasive procedures desires are more likely to consider ETOVA at the urging of their physicians/nurses. Providing appropriate healthcare education, medical insurance coverage, and information options for surgical treatments is vital to improving patients' acceptance of ETOVA.

Protective activity of tert-butylhydroquinone against oxidative stress and apoptosis induced by glutamate agonizts in R28 cells and mice retina.

Glutamate excitotoxicity can cause cell damage and apoptosis and play an important role in a variety of retinal diseases. Tertiary-butylhydroquinone (tBHQ) is an approved food-grade phenolic antioxidant with antioxidant activity in a variety of cells and tissues. We observed the protective effect of tBHQ on glutamatergic agonist-induced retina and explored its possible mechanism of action through in vitro cell experiments. The results showed that tBHQ had protective effects on NMDA-induced mouse retinal excitotoxicity and glutamate-induced excitotoxicity in rat retinal precursor cells (R28 cells). tBHQ reversed glutamate-induced apoptosis, production of intracellular reactive oxygen species, and reduction of mitochondrial membrane potential. Western blot analysis showed that tBHQ could increase the expression of procaspase-3, Bcl-2, AIF precursor, CAT, SOD2, Nrf2, NQO1, HO-1 and NF-κB in glutamate-treated cells, and decrease the expression of AIF cleavage products. Furthermore, we discovered that tBHQ activated müller glial cells. Based on these results, tBHQ may have antioxidant and anti-apoptotic properties, thus serving as a potential retinal protective agent. Its anti-oxidative stress effect was attributed to up-regulation of Nrf2, and its anti-apoptotic effect was related to its up-regulation of Bcl-2 expression and inhibition of mitochondria-dependent apoptosis.

PTB: Not just a polypyrimidine tract-binding protein.

Polypyrimidine tract-binding protein (PTB), as a member of the heterogeneous nuclear ribonucleoprotein family, functions by rapidly shuttling between the nucleus and the cytoplasm. PTB is involved in the alternative splicing of pre-messenger RNA (mRNA) and almost all steps of mRNA metabolism. PTB regulation is organ-specific; brain- or muscle-specific microRNAs and long noncoding RNAs partially contribute to regulating PTB, thereby modulating many physiological and pathological processes, such as embryonic development, cell development, spermatogenesis, and neuron growth and differentiation. Previous studies have shown that PTB knockout can inhibit tumorigenesis and development. The knockout of PTB in glial cells can be reprogrammed into functional neurons, which shows great promise in the field of nerve regeneration but is controversial.

Predicting the immune landscape of invasive breast carcinoma based on the novel signature of immune-related lncRNA.

BACKGROUND: The composition of the population of immune-related long non-coding ribonucleic acid (irlncRNA) generates a signature, irrespective of expression level, with potential value in predicting the survival status of patients with invasive breast carcinoma. METHODS: The current study uses univariate analysis to identify differentially expressed irlncRNA (DEirlncRNA) pairs from RNA-Seq data from The Cancer Genome Atlas (TCGA). 36 pairs of DEirlncRNA pairs were identified. Using various algorithms to construct a model, we have compared the area under the curve and calculated the 5-year curve of Akaike information criterion (AIC) values, which allows determination of the threshold indicating the maximum value for differentiation. Through cut-off point to establish the optimal model for distinguishing high-risk or low-risk groups among breast cancer patients. We assigned individual patients with invasive breast cancer to either high risk or low risk groups depending on the cut-off point, re-evaluated the tumor immune cell infiltration, the effectiveness of chemotherapy, immunosuppressive biomarkers, and immunotherapy. RESULTS: After re-assessing patients according to the threshold, we demonstrated an effective means of distinguish the severity of the disease, and identified patients with different clinicopathological characteristics, specific tumor immune infiltration states, high sensitivity to chemotherapy,wellpredicted response to immunotherapy and thus a more favorable survival outcome. CONCLUSIONS: The current study presents novel findings regarding the use of irlncRNA without the need to predict precise expression levels in the prognosis of breast cancer patients and to indicate their suitability for anti-tumor immunotherapy.

Deep Learning with Quantitative Features of Magnetic Resonance Images to Predict Biochemical Recurrence of Radical Prostatectomy: A Multi-Center Study.

Biochemical recurrence (BCR) occurs in up to 27% of patients after radical prostatectomy (RP) and often compromises oncologic survival. To determine whether imaging signatures on clinical prostate magnetic resonance imaging (MRI) could noninvasively characterize biochemical recurrence and optimize treatment. We retrospectively enrolled 485 patients underwent RP from 2010 to 2017 in three institutions. Quantitative and interpretable features were extracted from T2 delineated tumors. Deep learning-based survival analysis was then applied to develop the deep-radiomic signature (DRS-BCR). The model's performance was further evaluated, in comparison with conventional clinical models. The model achieved C-index of 0.802 in both primary and validating cohorts, outweighed the CAPRA-S score (0.677), NCCN model (0.586) and Gleason grade group systems (0.583). With application analysis, DRS-BCR model can significantly reduce false-positive predictions, so that nearly one-third of patients could benefit from the model by avoiding overtreatments. The deep learning-based survival analysis assisted quantitative image features from MRI performed well in prediction for BCR and has significant potential in optimizing systemic neoadjuvant or adjuvant therapies for prostate cancer patients.

Transplantation of reprogrammed peripheral blood cells differentiates into retinal ganglion cells in the mouse eye with NMDA-induced injury.

The generation of patient-specific induced pluripotent stem cells (iPSCs) holds significant implications for replacement therapy in treating optic neuropathies such as glaucoma. Stem-cell-based therapy targeted at replacing and replenishing retinal ganglion cells is progressing at a fast pace. However, clinical application necessitates an efficient and robust approach for cell manufacturing. Here, we examine whether the embryo body derived from human peripheral blood-derived iPSC can localize into the host retina and differentiate into retinal ganglion cells after transplantation into a glaucoma injury model. Human peripheral blood T cells were isolated and reprogrammed into an induced pluripotent stem cell (TiPSC) line using Sendai virus transduction carrying transcription factors Sox2, Klf4, c-Myc, and Oct4. TiPSCs were differentiated into RGC using neural basal culture. For in vivo studies, embryo bodies derived from TiPSCs (TiPSC-EB) were injected into the vitreous cavity of N-Methyl-d-aspartic acid (NMDA)-treated mice 2 weeks before sacrifice and retinal dissection. Induced pluripotent stem cells generated from human peripheral blood T cells display stem cell morphology and pluripotency markers. Furthermore, RGC-like cells differentiated from TiPSC exhibit extending axons and RGC marker TUJ1. When transplanted intravitreally into NMDA-treated mice, embryo bodies derived from TiPSC survived, migrated, and incorporated into the retina's GCL layer. In addition, TiPSC-EB transplants were able to differentiate into TUJ1 positive RGC-like cells. Retinal ganglion cells can be differentiated using human peripheral blood cells derived iPSC. Transplantation of embryo body derived from TiPSCs into a glaucoma mouse model could incorporate into host GCL and differentiate into RGC-like cells.

Patient-Level Prediction of Multi-Classification Task at Prostate MRI Based on End-to-End Framework Learning From Diagnostic Logic of Radiologists.

The grade groups (GGs) of Gleason scores (Gs) is the most critical indicator in the clinical diagnosis and treatment system of prostate cancer. End-to-end method for stratifying the patient-level pathological appearance of prostate cancer (PCa) in magnetic resonance (MRI) are of high demand for clinical decision. Existing methods typically employ a statistical method for integrating slice-level results to a patient-level result, which ignores the asymmetric use of ground truth (GT) and overall optimization. Therefore, more domain knowledge (e.g., diagnostic logic of radiologists) needs to be incorporated into the design of the framework. The patient-level GT is necessary to be logically assigned to each slice of a MRI to achieve joint optimization between slice-level analysis and patient-level decision-making. In this paper, we propose a framework (PCa-GGNet-v2) that learns from radiologists to capture signs in a separate two-dimensional (2-D) space of MRI and further associate them for the overall decision, where all steps are optimized jointly in an end-to-end trainable way. In the training phase, patient-level prediction is transferred from weak supervision to supervision with GT. An association route records the attentional slice for reweighting loss of MRI slices and interpretability. We evaluate our method in an in-house multi-center dataset (N = 570) and PROSTATEx (N = 204), which yields five-classification accuracy over 80% and AUC of 0.804 at patient-level respectively. Our method reveals the state-of-the-art performance for patient-level multi-classification task to personalized medicine.

Posterior pole retinotomy for treatment of recurrent macular hole retinal detachment in highly myopic eyes: a pilot study.

BACKGROUND: To investigate the feasibility and efficacy of posterior pole retinotomy to treat recurrent macular hole retinal detachment (MHRD) in highly myopic patients. METHODS: We performed a retrospective study and reviewed the medical records in our hospital between January 1, 2016 and December 31, 2018. Highly myopic patients who received posterior pole retinotomy with silicone oil tamponade for their recurrent MHRD after pars plana vitrectomy were included in the analysis. Postoperative retinal reattachment, best-corrected visual acuity (BCVA), macular hole closure, and complications were evaluated. RESULTS: There were 11 patients (11 eyes) included in this study. All retinas were reattached. Silicone oil was successfully removed from all eyes 1.5-3 months after the surgery. Macular holes were completely closed in three eyes and remained flat open in eight eyes. The BCVA of all eyes improved significantly at 12 months after surgery (logarithm of the minimal angle of resolution, pre vs. postoperatively, 1.87 ± 0.44 vs. 1.15 ± 0.24, P < 0.05). None of the patients had complications such as endophthalmitis, fundus hemorrhage, retinal redetachment, and proliferative vitreoretinopathy. CONCLUSION: Posterior pole retinotomy is a safe and effective surgery to treat recurrent MHRD after pars plana vitrectomy in highly myopic patients.

CXC Chemokines as Therapeutic Targets and Prognostic Biomarkers in Skin Cutaneous Melanoma Microenvironment.

BACKGROUND: Skin Cutaneous Melanoma (SKCM) is a tumor of the epidermal melanocytes induced by gene activation or mutation. It is the result of the interaction between genetic, constitutional, and environmental factors. SKCM is highly aggressive and is the most threatening skin tumor. The incidence of the disease is increasing year by year, and it is the main cause of death in skin tumors around the world. CXC chemokines in the tumor microenvironment can regulate the transport of immune cells and the activity of tumor cells, thus playing an anti-tumor immunological role and affecting the prognosis of patients. However, the expression level of CXC chemokine in SKCM and its effect on prognosis are still unclear. METHOD: Oncomine, UALCAN, GEPIA, STRING, GeneMANIA, cBioPortal, TIMER, TRRUST, DAVID 6.8, and Metascape were applied in our research. RESULT: The transcription of CXCL1, CXCL5, CXCL8, CXCL9, CXCL10, and CXCL13 in SKCM tissues were significantly higher than those in normal tissues. The pathological stage of SKCM patients is closely related to the expression of CXCL4, CXCL9, CXCL10, CXCL11, CXCL12, and CXCL13. The prognosis of SKCM patients with low transcription levels of CXCL4, CXCL9, CXCL10, CXCL11, and CXCL13 is better. The differential expression of CXC chemokines is mainly associated with inflammatory response, immune response, and cytokine mediated signaling pathways. Our data indicate that the key transcription factors of CXC chemokines are RELA, NF-κB1 and SP1. The targets of CXC chemokines are mainly LCK, LYN, SYK, MAPK2, MAPK12, and ART. The relationship between CXC chemokine expression and immune cell infiltration in SKCM was closed. CONCLUSIONS: Our research provides a basis for screening SKCM biomarkers, predicting prognosis, and choosing immunotherapy.

Endoscopic Lateral Neck Dissection (IIA, IIB, III, and IV) Using a Breast Approach: Outcomes From a Series of the First 24 Cases.

INTRODUCTION: In order to avoid large neck scar caused by conventional lateral neck dissection. We have explored and introduced endoscopic lateral neck dissection (IIA, IIB, III, and IV) using a breast approach. Now, we summarized and shared the outcomes of the first 24 cases. MATERIALS AND METHODS: All the patients were treated in our institute from January 2017 to May 2018, and followed-up for more than 1 year. The details of patients and this technique have been summarized and analyzed retrospectively. RESULTS: A series of first 24 cases were successfully managed with this technique, and no cases were converted to an open approach. Among these 24 patients, levels III + IV dissection had been performed in 6 patients and levels II+III+IV dissection had been performed in 18 patients. The mean age, body mass index, and sex were 39.3±10.5 years old, 24.1±3.5, and 2 male/22 female, respectively. The average operative time of total operation and lateral neck dissection was 238.8±37.2 minutes and 128.8±21.1 minutes, respectively. The mean dissected lateral lymph nodes were 5.9±2.2 (level II) in 18 cases and 15.9±3.9 (levels III+IV) in 24 cases. In addition, with no severe complications to date, such as asphyxia, main nerves injury (cervical plexus, vagus nerve, etc.), and permanent hypoparathyroidism, nor permanent recurrent laryngeal nerve injury, and so on. However, unexpectedly, had some mild and common complications like transient hypocalcemia in 4 cases (16.67%), transient horse 1 case (4.2%), controllable lymphatic leakage in 2 cases (8.3%), and controllable jugular vein injury in 2 cases (8.3%). One year after the operation, 1 case found lung metastasis but no local recurrence. In other 23 patients, no recurrence/metastasis and the average of serum thyroglobulin is 3.2±3.8 ng/mL. CONCLUSIONS: This technique can yield adequate oncological dissection for selected patients. Endoscopic thyroidectomy along with lateral neck dissection using a breast approach may provide an option for selected patients who favor avoiding a visible neck incision.

Radiologist-like artificial intelligence for grade group prediction of radical prostatectomy for reducing upgrading and downgrading from biopsy.

Rationale: To reduce upgrading and downgrading between needle biopsy (NB) and radical prostatectomy (RP) by predicting patient-level Gleason grade groups (GGs) of RP to avoid over- and under-treatment. Methods: In this study, we retrospectively enrolled 575 patients from two medical institutions. All patients received prebiopsy magnetic resonance (MR) examinations, and pathological evaluations of NB and RP were available. A total of 12,708 slices of original male pelvic MR images (T2-weighted sequences with fat suppression, T2WI-FS) containing 5405 slices of prostate tissue, and 2,753 tumor annotations (only T2WI-FS were annotated using RP pathological sections as ground truth) were analyzed for the prediction of patient-level RP GGs. We present a prostate cancer (PCa) framework, PCa-GGNet, that mimics radiologist behavior based on deep reinforcement learning (DRL). We developed and validated it using a multi-center format. Results: Accuracy (ACC) of our model outweighed NB results (0.815 [95% confidence interval (CI): 0.773-0.857] vs. 0.437 [95% CI: 0.335-0.539]). The PCa-GGNet scored higher (kappa value: 0.761) than NB (kappa value: 0.289). Our model significantly reduced the upgrading rate by 27.9% (P < 0.001) and downgrading rate by 6.4% (P = 0.029). Conclusions: DRL using MRI can be applied to the prediction of patient-level RP GGs to reduce upgrading and downgrading from biopsy, potentially improving the clinical benefits of prostate cancer oncologic controls.

Multiparametric MRI-based radiomics analysis for the prediction of breast tumor regression patterns after neoadjuvant chemotherapy.

OBJECTIVES: Breast cancers show different regression patterns after neoadjuvant chemotherapy. Certain regression patterns are associated with more reliable margins in breast-conserving surgery. Our study aims to establish a nomogram based on radiomic features and clinicopathological factors to predict regression patterns in breast cancer patients. METHODS: We retrospectively reviewed 144 breast cancer patients who received neoadjuvant chemotherapy and underwent definitive surgery in our center from January 2016 to December 2019. Tumor regression patterns were categorized as type 1 (concentric regression + pCR) and type 2 (multifocal residues + SD + PD) based on pathological results. We extracted 1158 multidimensional features from 2 sequences of MRI images. After feature selection, machine learning was applied to construct a radiomic signature. Clinical characteristics were selected by backward stepwise selection. The combined prediction model was built based on both the radiomic signature and clinical factors. The predictive performance of the combined prediction model was evaluated. RESULTS: Two radiomic features were selected for constructing the radiomic signature. Combined with two significant clinical characteristics, the combined prediction model showed excellent prediction performance, with an area under the receiver operating characteristic curve of 0.902 (95% confidence interval 0.8343-0.9701) in the primary cohort and 0.826 (95% confidence interval 0.6774-0.9753) in the validation cohort. CONCLUSIONS: Our study established a unique model combining a radiomic signature and clinicopathological factors to predict tumor regression patterns prior to the initiation of NAC. The early prediction of type 2 regression offers the opportunity to modify preoperative treatments or aids in determining surgical options.

Radiomics Analysis of Postoperative Epilepsy Seizures in Low-Grade Gliomas Using Preoperative MR Images.

Purpose: The present study aimed to evaluate the performance of radiomics features in the preoperative prediction of epileptic seizure following surgery in patients with LGG. Methods: This retrospective study collected 130 patients with LGG. Radiomics features were extracted from the T2-weighted MR images obtained before surgery. Multivariable Cox-regression with two nested leave-one-out cross validation (LOOCV) loops was applied to predict the prognosis, and elastic net was used in each LOOCV loop to select the predictive features. Logistic models were then built with the selected features to predict epileptic seizures at two time points. Student's t-tests were then used to compare the logistic model predicted probabilities of developing epilepsy in the epilepsy and non-epilepsy groups. The t-test was used to identify features that differentiated patients with early-onset epilepsy from their late-onset counterparts. Results: Seventeen features were selected with the two nested LOOCV loops. The index of concordance (C-index) of the Cox model was 0.683, and the logistic model predicted probabilities of seizure were significantly different between the epilepsy and non-epilepsy groups at each time point. Moreover, one feature was found to be significantly different between the patients with early- or late-onset epilepsy. Conclusion: A total of 17 radiomics features were correlated with postoperative epileptic seizures in patients with LGG and one feature was a significant predictor of the time of epilepsy onset.