-
IEEE Transactions on Pattern Analysis... Oct 2023Recent advances in deep learning have witnessed many successful unsupervised image-to-image translation models that learn correspondences between two visual domains...
Recent advances in deep learning have witnessed many successful unsupervised image-to-image translation models that learn correspondences between two visual domains without paired data. However, it is still a great challenge to build robust mappings between various domains especially for those with drastic visual discrepancies. In this paper, we introduce a novel versatile framework, Generative Prior-guided UNsupervised Image-to-image Translation (GP-UNIT), that improves the quality, applicability and controllability of the existing translation models. The key idea of GP-UNIT is to distill the generative prior from pre-trained class-conditional GANs to build coarse-level cross-domain correspondences, and to apply the learned prior to adversarial translations to excavate fine-level correspondences. With the learned multi-level content correspondences, GP-UNIT is able to perform valid translations between both close domains and distant domains. For close domains, GP-UNIT can be conditioned on a parameter to determine the intensity of the content correspondences during translation, allowing users to balance between content and style consistency. For distant domains, semi-supervised learning is explored to guide GP-UNIT to discover accurate semantic correspondences that are hard to learn solely from the appearance. We validate the superiority of GP-UNIT over state-of-the-art translation models in robust, high-quality and diversified translations between various domains through extensive experiments.
PubMed: 37289604
DOI: 10.1109/TPAMI.2023.3284003 -
BJS Open Nov 2023Tumour deposits are suggested to impact prognosis in colon cancer negatively. This study assessed the impact of tumour deposits on oncological outcomes.
BACKGROUND
Tumour deposits are suggested to impact prognosis in colon cancer negatively. This study assessed the impact of tumour deposits on oncological outcomes.
METHODS
Data from the Swedish Colorectal Cancer Registry for patients who underwent R0 abdominal surgery for TNM stage I-III colon cancer between 2011 and 2014 with 5-year follow-up were analysed with multivariable analysis. Patients were categorized for their tumour deposit status and compared for the local recurrence and distant metastasis rates and 5-year survivals (overall and relative). Subgroup analyses were performed according to the nodal disease status.
RESULTS
Of 8146 stage I-III colon cancer patients who underwent R0 resection, 8014 patients were analysed (808 tumour deposits positive, 7206 tumour deposits negative). Patients with tumour deposits positive tumours had increased local recurrence and distant metastasis rates (7.2 versus 3.0 per cent; P < 0.001 and 33.9 versus 12.0 per cent; P < 0.001 respectively) and reduced 5-year overall and relative survival (56.8 per cent versus 74.9 per cent; P < 0.001 and 68.5 versus 92.6 per cent; P < 0.001 respectively). In multivariable analysis, tumour deposits moderately increased the risks of local recurrence and distant metastasis (hazard ratio 1.50, 95 per cent c.i. 1.09 to 2.07; P = 0.013 and HR 1.91, 95 per cent c.i. 1.64 to 2.23; P < 0.001 respectively) and worse 5-year overall and relative survival (hazard ratio 1.60, 95 per cent c.i. 1.40 to 1.82; P < 0.001 and excess hazard ratio 2.24, 95 per cent c.i. 1.81 to 2.78; P < 0.001 respectively). Subgroup analysis of N stages found that N1c patients had worse outcomes than N0 for distant metastasis and relative survival. For patients with lymph node metastases tumour deposits increased the risks of distant metastasis and worse overall and relative survival, except for N2b patients.
CONCLUSION
Tumour deposits negatively impact the prognosis in colon cancer and must be considered when discussing adjuvant chemotherapy.
Topics: Humans; Extranodal Extension; Retrospective Studies; Colonic Neoplasms; Prognosis; Proportional Hazards Models
PubMed: 38035752
DOI: 10.1093/bjsopen/zrad122 -
Science (New York, N.Y.) May 2024Reservoirs of neutral hydrogen may block ionizing radiation from escaping distant galaxies.
Reservoirs of neutral hydrogen may block ionizing radiation from escaping distant galaxies.
PubMed: 38781396
DOI: 10.1126/science.adp5153 -
Frontiers in Oncology 2024Extraocular sebaceous carcinoma (SC), particularly those outside the head and neck region, is rare and not well-described.
INTRODUCTION
Extraocular sebaceous carcinoma (SC), particularly those outside the head and neck region, is rare and not well-described.
PURPOSE
This study aimed to explore the epidemiology and identify the prognostic factors of non-head and neck SC, describe the possible relevant factors of distant metastasis, and provide implications for distant metastasis screening.
METHODS
Data from the 17 registries in the Surveillance, Epidemiology, and End Results database were retrospectively collected for patients with SC outside the head and neck from 2000 through 2020. Overall survival (OS) and disease-specific survival (DSS) were the primary endpoints. Survival analysis was conducted through Kaplan-Meier curves, and multivariate analysis was carried out using Cox proportional hazard models.
RESULTS
A total of 1,237 patients with SC outside the head and neck were identified. The mean age at diagnosis of the entire patient cohort was 67.7 years (30 to 90+ years), and the mean tumor size was 2.2 cm (0.1-16 cm). Patients with distant disease experienced the lowest OS (mean, 29.5 months) than those with localized disease and regional disease ( < 0.0001). Multivariate analysis revealed that age, tumor size, and stage were independent determinants of OS; age, stage, and primary site were independent determinants of DSS. Tumor grade and lymph node status had less prognostic value for survival. Undifferentiated tumors have a trend toward distant metastasis, especially those at the primary site of the trunk.
CONCLUSION
The prognosis of the non-head and neck SC is excellent, while the survival of distant disease is very poor. Distant metastasis screening can be considered for undifferentiated tumors, especially those located in the trunk region with large tumor sizes.
PubMed: 38800410
DOI: 10.3389/fonc.2024.1395273 -
Surgery Jan 2024Parathyroid carcinoma is a rare malignancy with high recurrence rates. Liquid biopsy is a stratifying tool in disease recurrence/progression in other malignant...
BACKGROUND
Parathyroid carcinoma is a rare malignancy with high recurrence rates. Liquid biopsy is a stratifying tool in disease recurrence/progression in other malignant processes. This study sought to assess the feasibility and application of liquid biopsy in parathyroid carcinoma and its impact on surveillance.
METHODS
Retrospective review of a prospectively maintained database of adults treated for parathyroid carcinoma at a tertiary care center (2017-2023). Demographics, clinical characteristics, and laboratory variables were collected. Circulating cell-free deoxyribonucleic acid enrichment and circulating tumor cell enumeration were obtained from serial blood samples.
RESULTS
A total of 25 patients were identified-64% were male patients, with a median age of 56 years (interquartile range 45-63). Fifty blood samples were collected postoperatively. At first, circulating tumor cell enumeration, 56% (14/25) of patients had no evidence of disease, and 32% (8/25) had distant metastasis. Median follow-up was 53 months (interquartile range 23-107). At the last follow-up, 40% (10/25) of patients were found to have distant metastasis. Serial circulating tumor cell enumeration was performed in 52% of patients, median highest circulating tumor cell was (interquartile range 1-22). Circulating cell-free deoxyribonucleic acid was assessed in 64% of patients (16/25). There was no difference in circulating tumor cells or circulating cell-free deoxyribonucleic acid between those with distant metastasis and those without distant metastasis. The most common mutation identified was TP53, present in 88% of circulating cell-free deoxyribonucleic acid samples with a mutation. Circulating cell-free deoxyribonucleic acid and parathyroid hormone levels were not found to have any association (r = -0.27, P = .39), but parathyroid hormone and circulating tumor cell had a linear relationship (r = 0.76, P < .001).
CONCLUSION
Liquid biopsy appears to be a feasible tool in parathyroid carcinoma surveillance. The relationship between circulating cell-free deoxyribonucleic acid and parathyroid hormone levels remains unclear, and the association between circulating tumor cell enumeration and parathyroid hormone levels may be impactful. The finding that TP53 mutation is more prevalent in patients with distant metastasis may impact further management.
Topics: Adult; Humans; Male; Middle Aged; Female; Neoplastic Cells, Circulating; Parathyroid Neoplasms; Neoplasm Recurrence, Local; Liquid Biopsy; Parathyroid Hormone; Cell-Free Nucleic Acids
PubMed: 37993289
DOI: 10.1016/j.surg.2023.07.043 -
Frontiers in Surgery 2024Due to the novel advanced screening methods, the number of patients diagnosed with stage I colorectal cancer (CRC) is increasing. This retrospective cohort study aimed...
BACKGROUND
Due to the novel advanced screening methods, the number of patients diagnosed with stage I colorectal cancer (CRC) is increasing. This retrospective cohort study aimed to identify recurrence and survival risk factors of patients with stage I CRC after surgery.
MATERIALS AND METHODS
Patients with stage I CRC were evaluated, and their demographic and clinicopathologic variables were recorded. The log-rank test assessed the association of variables with overall survival (OS), recurrence-free survival (RFS), local recurrence, and distant metastasis.
RESULTS
The median overall survival period was 51 months. The recurrence rate was 13.7%: 7.2% local and 9.3% distant recurrence. One-, two-, three-, and five-year RFS were 92%, 89%, 87%, and 83%, respectively, and OS were 96%, 93%, 90%, and 89%, respectively. Local and distant recurrence rates were higher in patients with higher tumor grades. Additionally, RFS and OS were worse in patients with higher tumor grades, and perforation was associated with worse OS.
CONCLUSIONS
The determinants of survival and recurrence identified in the present study can be used to improve patient outcomes by early diagnosis and appropriate management of high-risk patients.
PubMed: 38817946
DOI: 10.3389/fsurg.2024.1377733 -
Gynecologic Oncology Aug 2023Treatment for endometrial cancer (EC) is increasingly guided by molecular risk classifications. Here, we aimed at using machine learning (ML) to incorporate clinical and...
INTRODUCTION
Treatment for endometrial cancer (EC) is increasingly guided by molecular risk classifications. Here, we aimed at using machine learning (ML) to incorporate clinical and molecular risk factors to optimize risk assessment.
METHODS
The Cancer Genome Atlas-Uterine Corpus Endometrial Carcinoma (n = 596), Memorial Sloan Kettering-Metastatic Events and Tropisms (n = 1315) and the American Association for Cancer Research Project Genomics Evidence Neoplasia Information Exchange (n = 4561) datasets were used to identify genetic alterations and clinicopathological features. Software packages including Keras, Pytorch, and Scikit Learn were tested to build artificial neural networks (ANNs) with a binary output as either intra-abdominal metastatic progression ('1') vs. non-metastatic ('0').
RESULTS
Black patients with EC have worse prognosis than White patients, adjusting for TP53 or POLE mutation status. Over 75% of Black patients carry TP53 mutations as compared to approximately 40% of White patients. Older age is associated with an increasing likelihood of TP53 mutation, high risk histology, and distant metastasis. For patients above age 70, 91% of Black and 60% of White EC patients carry TP53 mutations. A ML-based New Unified classifiCATion Score (NU-CATS) that incorporates age, race, histology, mismatch repair status, and TP53 mutation status showed 75% accuracy in prognosticating intra-abdominal progression. A higher NU-CATS is associated with an increasing risk of having positive pelvic or para-aortic lymph nodes and distant metastasis. NU-CATS was shown to outperform Leiden/TransPORTEC model for estimating risk of FIGO Stage I/II disease progression and survival in Black EC patients.
CONCLUSION
The NU-CATS, a ML-based, cost-effective algorithm, incorporates diverse clinicopathologic and molecular variables of EC and yields superior prognostication of the risk of nodal involvement, distant metastasis, disease progression, and overall survival.
Topics: Humans; Female; Cost-Benefit Analysis; Endometrial Neoplasms; Prognosis; Risk Factors; Mutation; Disease Progression
PubMed: 37336081
DOI: 10.1016/j.ygyno.2023.06.008 -
IScience Sep 2023The evolutionary trajectories of genomic alterations underlying distant recurrence in glioma remain largely unknown. To elucidate glioma evolution, we analyzed the...
The evolutionary trajectories of genomic alterations underlying distant recurrence in glioma remain largely unknown. To elucidate glioma evolution, we analyzed the evolutionary trajectories of matched pairs of primary tumors and relapse tumors or tumor fluid (TISF) based on deep whole-genome sequencing data (ctDNA). We found that MMR gene mutations occurred in the late stage in IDH-mutant glioma during gene evolution, which activates multiple signaling pathways and significantly increases distant recurrence potential. The proneural subtype characterized by PDGFRA amplification was likely prone to hypermutation and distant recurrence following treatment. The classical and mesenchymal subtypes tended to progress locally through subclonal reconstruction, trunk genes transformation, and convergence evolution. EGFR and NOTCH signaling pathways and CDNK2A mutation play an important role in promoting tumor local progression. Glioma subtypes displayed distinct preferred evolutionary patterns. ClinicalTrials.gov, NCT05512325.
PubMed: 37649695
DOI: 10.1016/j.isci.2023.107528 -
Palliative & Supportive Care May 2024
PubMed: 38695375
DOI: 10.1017/S1478951524000737 -
Endocrine Jun 2024Distant metastasis of thyroid cancer often indicates poor prognosis, and it is important to identify patients who have developed distant metastasis or are at high risk...
OBJECTIVE
Distant metastasis of thyroid cancer often indicates poor prognosis, and it is important to identify patients who have developed distant metastasis or are at high risk as early as possible. This paper aimed to predict distant metastasis of thyroid cancer through the construction of machine learning models to provide a reference for clinical diagnosis and treatment.
MATERIALS & METHODS
Data on demographic and clinicopathological characteristics of thyroid cancer patients between 2010 and 2015 were extracted from the National Institutes of Health (NIH) Surveillance, Epidemiology, and End Results (SEER) database. Our research used univariate and multivariate logistic models to screen independent risk factors, respectively. Decision Trees (DT), ElasticNet (ENET), Logistic Regression (LR), Extreme Gradient Boosting (XGBoost), Random Forest (RF), Multilayer Perceptron (MLP), Radial Basis Function Support Vector Machine (RBFSVM) and seven machine learning models were compared and evaluated by the following metrics: the area under receiver operating characteristic curve (AUC), calibration curve, decision curve analysis (DCA), sensitivity(also called recall), specificity, precision, accuracy and F1 score. Interpretable machine learning was used to identify possible correlation between variables and distant metastasis.
RESULTS
Independent risk factors for distant metastasis, including age, gender, race, marital status, histological type, capsular invasion, and number of lymph nodes metastases were screened by multifactorial regression analysis. Among the seven machine learning algorithms, RF was the best algorithm, with an AUC of 0.948, sensitivity of 0.919, accuracy of 0.845, and F1 score of 0.886 in the training set, and an AUC of 0.960, sensitivity of 0.929, accuracy of 0.906, and F1 score of 0.908 in the test set.
CONCLUSIONS
The machine learning model constructed in this study helps in the early diagnosis of distant thyroid metastases and helps physicians to make better decisions and medical interventions.
Topics: Humans; Thyroid Neoplasms; Female; Machine Learning; Male; SEER Program; Middle Aged; Adult; Aged; Risk Factors; Prognosis; Neoplasm Metastasis; Databases, Factual
PubMed: 38155324
DOI: 10.1007/s12020-023-03657-4