Prediction Of Therapeutic Response Research Articles

Metabolomic biomarkers in liquid biopsy: accurate cancer diagnosis and prognosis monitoring.

Liquid biopsy, a novel detection method, has recently become an active research area in clinical cancer owing to its unique advantages. Studies on circulating free DNA, circulating tumor cells, and exosomes obtained by liquid biopsy have shown great advances and they have entered clinical practice as new cancer biomarkers. The metabolism of the body is dynamic as cancer originates and progresses. Metabolic abnormalities caused by cancer can be detected in the blood, sputum, urine, and other biological fluids via systemic or local circulation. A considerable number of recent studies have focused on the roles of metabolic molecules in cancer. The purpose of this review is to provide an overview of metabolic markers from various biological fluids in the latest clinical studies, which may contribute to cancer screening and diagnosis, differentiation of cancer typing, grading and staging, and prediction of therapeutic response and prognosis.

AAV gene therapy in companion dogs with severe hemophilia: Real-world long-term data on immunogenicity, efficacy, and quality of life

The hemophilias are the most common severe inherited bleeding disorders and are caused by deficiency of clotting factor (F) VIII (hemophilia A) or FIX (hemophilia B). The resultant bleeding predisposition significantly increases morbidity and mortality. The ability to improve the bleeding phenotype with modest increases in clotting factor levels has enabled the development and regulatory approval of adeno-associated viral (AAV) vector gene therapies for people with hemophilia A and B. The canine hemophilia model has proven to be one of the best predictors of therapeutic response in humans. Here, we report long-term follow-up of 12 companion dogs with severe hemophilia that were treated in a real-world setting with AAV gene therapy. Despite more baseline bleeding than in research dogs, companion dogs demonstrated a 94% decrease in bleeding rates and 61% improvement in quality-of-life over a median of 4.1 years (range 2.6-8.9). No new anti-transgene immune responses were detected; 1 dog with a preexisting anti-FVIII inhibitor achieved immune tolerance with gene therapy. Two dogs expressing 1-5% FVIII post gene therapy experienced fatal bleeding events. These data suggest AAV liver-directed gene therapy is efficacious in a real-world setting, but should target expression > 5% and closely monitor those with levels in the 1-5% range.

Histopathology images-based deep learning prediction of prognosis and therapeutic response in small cell lung cancer

Small cell lung cancer (SCLC) is a highly aggressive subtype of lung cancer characterized by rapid tumor growth and early metastasis. Accurate prediction of prognosis and therapeutic response is crucial for optimizing treatment strategies and improving patient outcomes. In this study, we conducted a deep-learning analysis of Hematoxylin and Eosin (H&E) stained histopathological images using contrastive clustering and identified 50 intricate histomorphological phenotype clusters (HPCs) as pathomic features. We identified two of 50 HPCs with significant prognostic value and then integrated them into a pathomics signature (PathoSig) using the Cox regression model. PathoSig showed significant risk stratification for overall survival and disease-free survival and successfully identified patients who may benefit from postoperative or preoperative chemoradiotherapy. The predictive power of PathoSig was validated in independent multicenter cohorts. Furthermore, PathoSig can provide comprehensive prognostic information beyond the current TNM staging system and molecular subtyping. Overall, our study highlights the significant potential of utilizing histopathology images-based deep learning in improving prognostic predictions and evaluating therapeutic response in SCLC. PathoSig represents an effective tool that aids clinicians in making informed decisions and selecting personalized treatment strategies for SCLC patients.

Biomarkers-Aware Asymmetric Bibranch GAN with Adaptive Memory Batch Normalization for Prediction of Anti-VEGF Treatment Response in Neovascular Age-related Macular Degeneration.

The emergence of anti-vascular endothelial growth factor (anti-VEGF) therapy has revolutionized neovascular age-related macular degeneration (nAMD). Post-therapeutic optical coherence tomography (OCT) imaging facilitates the prediction of therapeutic response to anti-VEGF therapy for nAMD. Although the generative adversarial network (GAN) is a popular generative model for post-therapeutic OCT image generation, it is realistically challenging to gather sufficient pre- and post-therapeutic OCT image pairs, resulting in overfitting. Moreover, the available GAN-based methods ignore local details, such as the biomarkers that are essential for nAMD treatment. To address these issues, a Biomarkers-aware Asymmetric Bibranch GAN (BAABGAN) is proposed to efficiently generate post-therapeutic OCT images. Specifically, one branch is developed to learn prior knowledge with a high degree of transferability from large-scale data, termed the source branch. Then, the source branch transfer knowledge to another branch, which is trained on small-scale paired data, termed the target branch. To boost the transferability, a novel Adaptive Memory Batch Normalization (AMBN) is introduced in the source branch, which learns more effective global knowledge that is impervious to noise via memory mechanism. Also, a novel Adaptive Biomarkers-aware Attention (ABA) module is proposed to encode biomarkers information into latent features of target branches to learn finer local details of biomarkers. The proposed method outperforms traditional GAN models and can produce high-quality post-treatment OCT pictures with limited data sets, as shown by the results of experiments.

Ct-based subregional radiomics using hand-crafted and deep learning features for prediction of therapeutic response to anti-PD1 therapy in NSCLC

PurposeTo develop and externally validate subregional radiomics for predicting therapeutic response to anti-PD1 therapy in non-small-cell lung cancer (NSCLC). MethodsSixty-six patients from center 1 served as training and internal validation cohorts. Thirty patients from center 2 and thirty patients from center 3 served as external validation 1 and external validation 2 cohorts, respectively. The lesions identified on CT scans were subdivided into two phenotypically consistent subregions by automatic clustering on the patient-level and population-level (denoted as marginal S1 and inner S2). Handcrafted and deep learning-based features were extracted separately from the entire tumor region and subregions, then selected using the intraclass correlation coefficient and least absolute shrinkage and selection operator regression (LASSO). Radiomics signatures (RSs) were built integrating the selected features and correlation coefficients using a logistic regression method. Area under the receiver operating characteristic (ROC) curve (AUC) was calculated to assess the RSs. ResultsRSs derived from S1 outperformed those from S2 and the whole tumor region for both handcrafted and deep learning features. The Fusion-RS incorporating the two feature types achieved the best prediction performance in training (AUC = 0.947, 95 % Confidence Interval [CI] 0.905–0.989, SPE = 0.895, SEN = 0.878), internal validation (AUC = 0.875, 95 % CI: 0.782–0.969, SPE = 0.724, SEN = 0.952), external validation 1 (AUC = 0.836, 95 % CI: 0.694–0.977, SPE = 1.000, SEN = 0.533) and external validation 2 (AUC = 0.783, 95 % CI: 0.613–0.953, SPE = 0.765, SEN = 0.692) cohorts. ConclusionsSubregional radiomics analysis can be useful for predicting therapeutic response to anti-PD1 therapy. The developed Fusion-RS may be considered as a potential non-invasive tool for individual treatment managements.

Genomics in Clinical trials for Breast Cancer.

Breast cancer (B.C.) still has increasing incidences and mortality rates globally. It is known that B.C. and other cancers have a very high rate of genetic heterogeneity and genomic mutations. Traditional oncology approaches have not been able to provide a lasting solution. Targeted therapeutics have been instrumental in handling the complexity and resistance associated with B.C. However, the progress of genomic technology has transformed our understanding of the genetic landscape of breast cancer, opening new avenues for improved anti-cancer therapeutics. Genomics is critical in developing tailored therapeutics and identifying patients most benefit from these treatments. The next generation of breast cancer clinical trials has incorporated next-generation sequencing technologies into the process, and we have seen benefits. These innovations have led to the approval of better-targeted therapies for patients with breast cancer. Genomics has a role to play in clinical trials, including genomic tests that have been approved, patient selection and prediction of therapeutic response. Multiple clinical trials in breast cancer have been done and are still ongoing, which have applied genomics technology. Precision medicine can be achieved in breast cancer therapy with increased efforts and advanced genomic studies in this domain. Genomics studies assist with patient outcomes improvement and oncology advancement by providing a deeper understanding of the biology behind breast cancer. This article will examine the present state of genomics in breast cancer clinical trials.

Value of 68Ga-FAPI-04 and 18F-FDG PET/CT in Early Prediction of Pathologic Response to Neoadjuvant Chemotherapy in Locally Advanced Gastric Cancer.

This prospective study investigated whether PET parameters from 18F-FDG and 68Ga-fibroblast activation protein inhibitor (FAPI)-04 PET/CT can predict a pathologic response to neoadjuvant chemotherapy (NAC) early in patients with locally advanced gastric cancer (LAGC). Methods: The study included 28 patients with LAGC who underwent 18F-FDG PET/CT and 68Ga-FAPI-04 PET/CT at baseline and after 1 cycle of NAC. PET parameters including SUV and tumor-to-background ratio (TBR), as well as the change rate of SUV and TBR, were recorded. Patients were classified as major or minor pathologic responders according to postoperative pathology findings. We compared the PET parameters between the 2 pathologic response groups and different treatment regimens and analyzed their predictive performance for tumor pathologic response. Results: Major pathologic responders had significantly lower 68Ga-FAPI change rates (percentage SUVmax [%SUVmax], percentage SUVpeak [%SUVpeak], and percentage TBR [%TBR]) than minor pathologic responders. Among the PET parameters, 68Ga-FAPI %SUVmax (area under the curve, 0.856; P = 0.009), %SUVpeak (area under the curve, 0.811; P = 0.022), and %TBR (area under the curve, 0.864; P = 0.007) were significant parameters for early prediction of pathologic response to NAC in LAGC; they had the same predictive accuracy of 89.29%, with the thresholds of decrease to at least 52.43%, 60.46%, and 52.96%, respectively. In addition, 68Ga-FAPI %SUVmax and %TBR showed significant differences between the different treatment regimens. Conclusion: In this preliminary study, 68Ga-FAPI-04 PET change rate parameters were preferable to 18F-FDG in predicting pathologic response to NAC at an early stage in LAGC. 68Ga-FAPI %SUVmax and %TBR may be better predictors of therapeutic response between different treatment regimens. These findings may help optimize the treatment for patients with LAGC.

Identifying the programmed cell death index of hepatocellular carcinoma for prognosis and therapy response improvement by machine learning: a bioinformatics analysis and experimental validation.

Despite advancements in hepatocellular carcinoma (HCC) treatments, the prognosis for patients remains suboptimal. Cumulative evidence suggests that programmed cell death (PCD) exerts crucial functions in HCC. PCD-related genes are potential predictors for prognosis and therapeutic responses. A systematic analysis of 14 PCD modes was conducted to determine the correlation between PCD and HCC. A novel machine learning-based integrative framework was utilized to construct the PCD Index (PCDI) for prognosis and therapeutic response prediction. A comprehensive analysis of PCDI genes was performed, leveraging data including single-cell sequencing and proteomics. GBA was selected, and its functions were investigated in HCC cell lines by in vitro experiments. Two PCD clusters with different clinical and biological characteristics were identified in HCC. With the computational framework, the PCDI was constructed, demonstrating superior prognostic predictive efficacy and surpassing previously published prognostic models. An efficient clinical nomogram based on PCDI and clinicopathological factors was then developed. PCDI was intimately associated with immunological attributes, and PCDI could efficaciously predict immunotherapy response. Additionally, the PCDI could predict the chemotherapy sensitivity of HCC patients. A multilevel panorama of PCDI genes confirmed its stability and credibility. Finally, the knockdown of GBA could suppress both the proliferative and invasive capacities of HCC cells. This study systematically elucidated the association between PCD and HCC. A robust PCDI was constructed for prognosis and therapy response prediction, which would facilitate clinical management and personalized therapy for HCC.

Altered interhemispheric functional connectivity in patients with obsessive-compulsive disorder and its potential in therapeutic response prediction.

The trajectory of voxel-mirrored homotopic connectivity (VMHC) after medical treatment in obsessive-compulsive disorder (OCD) and its value in prediction of treatment response remains unclear. This study aimed to investigate the pathophysiological mechanism of OCD, as well as biomarkers for prediction of pharmacological efficacy. Medication-free patients with OCD and healthy controls (HCs) underwent magnetic resonance imaging. The patients were scanned again after a 4-week treatment with paroxetine. The acquired data were subjected to VMHC, support vector regression (SVR), and correlation analyses. Compared with HCs (36 subjects), patients with OCD (34 subjects after excluding two subjects with excessive head movement) exhibited significantly lower VMHC in the bilateral superior parietal lobule (SPL), postcentral gyrus, and calcarine cortex, and VMHC in the postcentral gyrus was positively correlated with cognitive function. After treatment, the patients showed increased VMHC in the bilateral posterior cingulate cortex/precuneus (PCC/PCu) with the improvement of symptoms. SVR results showed that VMHC in the postcentral gyrus at baseline could aid to predict a change in the scores of OCD scales. This study revealed that SPL, postcentral gyrus, and calcarine cortex participate in the pathophysiological mechanism of OCD while PCC/PCu participate in the pharmacological mechanism. VMHC in the postcentral gyrus is a potential predictive biomarker of the treatment effects in OCD.

Disulfidptosis-Related lncRNA for the Establishment of Novel Prognostic Signature and Therapeutic Response Prediction to Endometrial Cancer.

Disulfidptosis, a newly discovered cellular death mechanism initiated by disulfide stress, features elevated expression of SLC7A11 and restricted glucose availability, rendering it a possible therapeutic target for cancer. Endometrial cancer of the uterine corpus (ECUC) ranks among prevalent gynecological malignancies. Long non-coding RNAs (lncRNAs) have been implicated in ECUC's metabolic pathways, invasive capabilities, and metastatic processes. Yet, the prognostic implications of Disulfidptosis-Linked lncRNAs (DLLs) in ECUC remain ambiguous. Transcriptome and clinical datasets related to ECUC were sourced from The Cancer Genome Atlas (TCGA), while genes linked with disulfidptosis were identified from existing literature. A panel of ten DLLs was discerned through least absolute shrinkage and selection operator (LASSO) coupled with Cox regression methods to formulate and validate risk prognostic models. We engineered a nomogram for ECUC patient prognosis forecasting and further examined the model via gene set enrichment analysis (GSEA), principal component analysis (PCA), gene set analysis (GSA), immune profiling, and sensitivity to antineoplastic agents. Prognostic models employing a set of ten DLLs (including AC005034.2, AC020765.2, AL158071.4, AL161663.2, AP000787.1, CR392039.3, EMSLR, SEC24B-AS1, Z69733.1, Z94721.3) were established. Based on median risk values, patient samples were stratified into high- and low-risk cohorts, revealing notable differences in survival across both training and validation datasets. The risk scores, when amalgamated with clinical variables, acted as standalone predictors of prognosis. GSEA findings indicated that the high-risk category predominantly aligned with pathways like extracellular matrix interactions and cell adhesion molecules, suggesting a likely association with metastatic potential. Concurrently, we scrutinized disparities in immune function and tumor mutational burden across risk categories and identified anticancer drugs with likely efficacy. In summary, a set of ten DLLs proved useful in forecasting patient outcomes and holds potential for informing targeted therapeutic approaches in ECUC.

Robust analysis of a novel PANoptosis-related prognostic gene signature model for hepatocellular carcinoma immune infiltration and therapeutic response

PANoptosis, an interplay between pyroptosis, apoptosis, and necroptosis, is deeply involved in cancer development and immunity. However, the influence of PANoptosis in hepatocellular carcinoma (HCC) remains to be further investigated. The differentially expressed PANoptosis-related genes (PANRGs) was screened in The Cancer Genome Atlas (TCGA) database. Accordingly, mutation, bioinformatics, and consensus clustering analyses were performed. Then, a prognostic risk model was developed by least absolute shrinkage and selection operator (LASSO) Cox regression. Furthermore, the prognostic value, immunity correlation and therapeutic response prediction ability of risk model were explored. A total of 18 PANRGs were differently expressed in the TCGA-HCC cohort and were mainly involved in cancer- and cell death-related signal pathways. Using unsupervised clustering method, we identified two PANRGs-mediated clustering patterns. The remarkable differences between the two clusters on overall survival (OS) and clinical features were demonstrated respectively. Based on the five-gene prognostic risk model, the calculated PANRG-scores were used to categorize the subgroups as high- and low-risk. Notably, the high-risk subgroup had a dismal prognosis and exhibited much lower immune infiltration levels of mast cells, nature killer cells and pDCs, but higher levels of aDCs, iDCs and Treg cells than those in the low-risk subgroup. Furthermore, we constructed a reliable nomogram combining clinical traits and PANRG-score to predict the OS of HCC patients. The significantly negative correlation between PANoptosis and tumor mutation burden (TMB), ferroptosis were revealed. In drug sensitivity analysis, the high-risk subgroup had a considerably lower TIDE score, suggesting a preferable response to immunotherapy, and may be more sensitive to Tipifarnib, Imatinib, Doxorubicin, and Gemcitabine. The upregulated mRNA expressions of FADD were validated in 16 paired HCC tissues of Guangxi cohort. Based on PANoptosis-related genes, an integrated risk signature was constructed to provide a roadmap for patient stratification and predict HCC patient's prognosis. The patients with the higher PANRG-score may carry a dismal survival and relatively low immune infiltration, but a potential better immunotherapy response. Therefore, future HCC therapy perspectives should emphasize the setting of PANoptosis to achieve a personalized, practicable and effective therapeutic regimen.

Deep learning model improves tumor-infiltrating lymphocyte evaluation and therapeutic response prediction in breast cancer

Tumor-infiltrating lymphocytes (TILs) have been recognized as key players in the tumor microenvironment of breast cancer, but substantial interobserver variability among pathologists has impeded its utility as a biomarker. We developed a deep learning (DL)-based TIL analyzer to evaluate stromal TILs (sTILs) in breast cancer. Three pathologists evaluated 402 whole slide images of breast cancer and interpreted the sTIL scores. A standalone performance of the DL model was evaluated in the 210 cases (52.2%) exhibiting sTIL score differences of less than 10 percentage points, yielding a concordance correlation coefficient of 0.755 (95% confidence interval [CI], 0.693–0.805) in comparison to the pathologists’ scores. For the 226 slides (56.2%) showing a 10 percentage points or greater variance between pathologists and the DL model, revisions were made. The number of discordant cases was reduced to 116 (28.9%) with the DL assistance (p < 0.001). The DL assistance also increased the concordance correlation coefficient of the sTIL score among every two pathologists. In triple-negative and human epidermal growth factor receptor 2 (HER2)-positive breast cancer patients who underwent the neoadjuvant chemotherapy, the DL-assisted revision notably accentuated higher sTIL scores in responders (26.8 ± 19.6 vs. 19.0 ± 16.4, p = 0.003). Furthermore, the DL-assistant revision disclosed the correlation of sTIL-high tumors (sTIL ≥ 50) with the chemotherapeutic response (odd ratio 1.28 [95% confidence interval, 1.01–1.63], p = 0.039). Through enhancing inter-pathologist concordance in sTIL interpretation and predicting neoadjuvant chemotherapy response, here we report the utility of the DL-based tool as a reference for sTIL scoring in breast cancer assessment.

Morphometry-based radiomics for predicting therapeutic response in patients with gliomas following radiotherapy.

Gliomas are still considered as challenging in oncologic management despite the developments in treatment approaches. The complete elimination of a glioma might not be possible even after a treatment and assessment of therapeutic response is important to determine the future course of actions for patients with such cancers. In the recent years radiomics has emerged as a promising solution with potential applications including prediction of therapeutic response. Hence, this study was focused on investigating whether morphometry-based radiomics signature could be used to predict therapeutic response in patients with gliomas following radiotherapy. 105 magnetic resonance (MR) images including segmented and non-segmented images were used to extract morphometric features and develop a morphometry-based radiomics signature. After determining the appropriate machine learning algorithm, a prediction model was developed to predict the therapeutic response eliminating the highly correlated features as well as without eliminating the highly correlated features. Then the model performance was evaluated. Tumor grade had the highest contribution to develop the morphometry-based signature. Random forest provided the highest accuracy to train the prediction model derived from the morphometry-based radiomics signature. An accuracy of 86% and area under the curve (AUC) value of 0.91 were achieved for the prediction model evaluated without eliminating the highly correlated features whereas accuracy and AUC value were 84% and 0.92 respectively for the prediction model evaluated after eliminating the highly correlated features. Nonetheless, the developed morphometry-based radiomics signature could be utilized as a noninvasive biomarker for therapeutic response in patients with gliomas following radiotherapy.

A novel disulfidptosis and glycolysis related risk score signature for prediction of prognosis and ICI therapeutic responsiveness in colorectal cancer

Disulfidptosis is a newly-identified non-programmed cell death mode with tight associations with glucose metabolism. Elevated glycolysis is an important metabolic feature of tumor cells, which fulfills the energy requirement for their rapid growth and progression. Our present study determined to develop a disulfidptosis and glycolysis related gene (DGRG) risk score signature to predict the prognosis and ICI therapeutic responsiveness for CRC patients. First, the gene expression and clinical profiles for CRC patients were obtained from TCGA and GEO database. Using weighted gene co-expression network analysis, we identified hub genes showing the strongest correlations with both disulfidptosis and glycolysis activities. Next, a DGRG risk score signature was successfully developed through univariate and least absolute shrinkage and selection operator method Cox regression method. A DGRG risk score-based nomogram could further enhance the predictive performance. In addition, an array of systemic analysis was performed to unravel the correlation of DGRG risk score with tumor microenvironment. The results showed that CRC patients with low DGRG risk level had up-regulated immune cell infiltrations, enhanced metabolic activities and heightened gene mutation frequencies, while high risk patients was the opposite. Moreover, our present study identified low risk CRC patients as potential beneficiaries from immune checkpoint inhibitor (ICI) therapies. Our present work highlighted the potential utility of DGRG risk score signature in prognosis prediction and ICI responsiveness determination for CRC patients, which demonstrated promising clinical application value.

Patient-Derived Ex Vivo Cultures and Endpoint Assays with Surrogate Biomarkers in Functional Testing for Prediction of Therapeutic Response.

Prediction of therapeutic outcomes is important for cancer patients in order to reduce side effects and improve the efficacy of anti-cancer drugs. Currently, the most widely accepted method for predicting the efficacy of anti-cancer drugs is gene panel testing based on next-generation sequencing. However, gene panel testing has several limitations. For example, only 10% of cancer patients are estimated to have druggable mutations, even if whole-exome sequencing is applied. Additionally, even if optimal drugs are selected, a significant proportion of patients derive no benefit from the indicated drug treatment. Furthermore, most of the anti-cancer drugs selected by gene panel testing are molecularly targeted drugs, and the efficacies of cytotoxic drugs remain difficult to predict. Apart from gene panel testing, attempts to predict chemotherapeutic efficacy using ex vivo cultures from cancer patients have been increasing. Several groups have retrospectively demonstrated correlations between ex vivo drug sensitivity and clinical outcome. For ex vivo culture, surgically resected tumor tissue is the most abundant source. However, patients with recurrent or metastatic tumors do not usually undergo surgery, and chemotherapy may be the only option for those with inoperable tumors. Therefore, predictive methods using small amounts of cancer tissue from diagnostic materials such as endoscopic, fine-needle aspirates, needle cores and liquid biopsies are needed. To achieve this, various types of ex vivo culture and endpoint assays using effective surrogate biomarkers of drug sensitivity have recently been developed. Here, we review the variety of ex vivo cultures and endpoint assays currently available.

Genetic Polymorphisms of rs9949644 in MAPK4 Are Associated with Clinical Response to Methotrexate in Patients with Psoriasis

Background: The study aimed to investigate the relationship of MAPK4 genetic variants with the efficacy of methotrexate (MTX) in psoriasis patients. Methods: Patients treated with MTX were classified as responders or nonresponders if the Psoriasis Area and Severity Index (PASI) at week 12 was reduced to greater than 75% or lower than 75%, respectively. The genotypes of 14 MAPK4 single-nucleotide polymorphisms in 310 patients were analyzed. The expression levels of MAPK4 protein were detected by Western blot. Results: Only rs9949644 polymorphisms were associated with the efficacy after adjusting for the confounding factors. Patients with the rs9949644 AG or GG genotype had a better clinical response compared to patients with the AA genotype. Rs9949644 polymorphisms were significantly associated with the PASI improvement rate. Besides, the protein level of MAPK4, positively associated with the psoriasis severity, was higher in patients. There were no significant differences of MAPK4 protein levels among the three groups. While after treatment, MAPK4 levels in the AG or GG group showed a significantly down-regulated trend. Conclusion: By demonstrating the significant association of MAPK4 with the efficacy of MTX, this study indicates that MAPK4 may be involved in the psoriasis progression and act as a predictor of therapeutic response.

Recent advances of pathomics in colorectal cancer diagnosis and prognosis.

Colorectal cancer (CRC) is one of the most common malignancies, with the third highest incidence and the second highest mortality in the world. To improve the therapeutic outcome, the risk stratification and prognosis predictions would help guide clinical treatment decisions. Achieving these goals have been facilitated by the fast development of artificial intelligence (AI) -based algorithms using radiological and pathological data, in combination with genomic information. Among them, features extracted from pathological images, termed pathomics, are able to reflect sub-visual characteristics linking to better stratification and prediction of therapeutic responses. In this paper, we review recent advances in pathological image-based algorithms in CRC, focusing on diagnosis of benign and malignant lesions, micro-satellite instability, as well as prediction of neoadjuvant chemoradiotherapy and the prognosis of CRC patients.

Learning Individual Survival Models from PanCancer Whole Transcriptome Data.

Personalized medicine attempts to predict survival time for each patient, based on their individual tumor molecular profile. We investigate whether our survival learner in combination with a dimension reduction method can produce useful survival estimates for a variety of patients with cancer. This article provides a method that learns a model for predicting the survival time for individual patients with cancer from the PanCancer Atlas: given the (16,335 dimensional) gene expression profiles from 10,173 patients, each having one of 33 cancers, this method uses unsupervised nonnegative matrix factorization (NMF) to reexpress the gene expression data for each patient in terms of 100 learned NMF factors. It then feeds these 100 factors into the Multi-Task Logistic Regression (MTLR) learner to produce cancer-specific models for each of 20 cancers (with >50 uncensored instances); this produces "individual survival distributions" (ISD), which provide survival probabilities at each future time for each individual patient, which provides a patient's risk score and estimated survival time. Our NMF-MTLR concordance indices outperformed the VAECox benchmark by 14.9% overall. We achieved optimal survival prediction using pan-cancer NMF in combination with cancer-specific MTLR models. We provide biological interpretation of the NMF model and clinical implications of ISDs for prognosis and therapeutic response prediction. NMF-MTLR provides many benefits over other models: superior model discrimination, superior calibration, meaningful survival time estimates, and accurate probabilistic estimates of survival over time for each individual patient. We advocate for the adoption of these cancer survival models in clinical and research settings.

Deep learning of 2D-Restructured gene expression representations for improved low-sample therapeutic response prediction

Clinical outcome prediction is important for stratified therapeutics. Machine learning (ML) and deep learning (DL) methods facilitate therapeutic response prediction from transcriptomic profiles of cells and clinical samples. Clinical transcriptomic DL is challenged by the low-sample sizes (34–286 subjects), high-dimensionality (up to 21,653 genes) and unordered nature of clinical transcriptomic data. The established methods rely on ML algorithms at accuracy levels of 0.6–0.8 AUC/ACC values. Low-sample DL algorithms are needed for enhanced prediction capability. Here, an unsupervised manifold-guided algorithm was employed for restructuring transcriptomic data into ordered image-like 2D-representations, followed by efficient DL of these 2D-representations with deep ConvNets. Our DL models significantly outperformed the state-of-the-art (SOTA) ML models on 82% of 17 low-sample benchmark datasets (53% with >0.05 AUC/ACC improvement). They are more robust than the SOTA models in cross-cohort prediction tasks, and in identifying robust biomarkers and response-dependent variational patterns consistent with experimental indications.

Optimal time for early therapeutic response prediction in nasopharyngeal carcinoma with functional magnetic resonance imaging.

Physiological changes in tumour occur much earlier than morphological changes. They can potentially be used as biomarkers for therapeutic response prediction. This study aimed to investigate the optimal time for early therapeutic response prediction with multi-parametric magnetic resonance imaging (MRI) in patients with nasopharyngeal carcinoma (NPC) receiving concurrent chemo-radiotherapy (CCRT). Twenty-seven NPC patients were divided into the responder (N=23) and the poor-responder (N=4) groups by their primary tumour post-treatment shrinkages. Single-voxel proton MR spectroscopy (1H-MRS), diffusion-weighted (DW) and dynamic contrast-enhanced (DCE) MRI were scanned at baseline, weekly during CCRT and post-CCRT. The median choline peak in 1H-MRS, the median apparent diffusion coefficient (ADC) in DW-MRI, the median influx rate constant (Ktrans), reflux rate constant (Kep), volume of extravascular-extracellular space per unit volume (Ve), and initial area under the time-intensity curve for the first 60s (iAUC60) in DCE-MRI were compared between the two groups with the Mann-Whitney tests for any significant difference at different time points. In DW-MRI, the percentage increase in ADC from baseline to week-1 for the responders (median=11.39%, IQR=18.13%) was higher than the poor-responders (median=4.91%, IQR=7.86%) (p=0.027). In DCE-MRI, the iAUC60 on week-2 was found significantly higher in the poor-responders (median=0.398, IQR=0.051) than the responders (median=0.192, IQR=0.111) (p=0.012). No significant difference was found in median choline peaks in 1H-MRS at all time points. Early perfusion and diffusion changes occurred in primary tumours of NPC patients treated with CCRT. The DW-MRI on week-1 and the DCE-MRI on week-2 were the optimal time points for early therapeutic response prediction.