Hasil untuk "Neoplasms. Tumors. Oncology. Including cancer and carcinogens"

Svetlana Tamkovich, Svetlana Tamkovich, Aleksei Shefer et al.

Breast cancer (BC) remains the most common malignant disease in women. However, currently used instrumental and laboratory (CA15-3, CA125, etc.) diagnostic methods demonstrate insufficient sensitivity and specificity for early and reliable detection of BC. In this regard, great expectations are associated with the liquid biopsy method based on the identification of tumor cells or their components, including tumor-derived exosomes. The purpose of this study is to analyze current data on exosome proteins that can be used for diagnostics using liquid biopsy. This review discusses the role of exosomal crown proteins in the spread of BC and assesses their potential as diagnostic markers. The undoubted advantages of using exosomal crown proteins as tumor markers compared to other components of the tumor secretome are the simplicity and reproducibility of their analysis by flow cytometry, as well as, unlike microRNA, tissue specificity. In contrast to prior reviews that primarily catalogue extracellular vesicle cargo, we specifically assess surface-accessible proteins that combine biological relevance with analytical feasibility. This approach bridges mechanistic EV biology with the practical design of clinically translatable diagnostic assays. Standardization of protocols for exosome isolation, antibody validation, and signal amplification will be critical to the successful implementation of this approach into routine clinical practice. Integration of exosomal coronary protein profiling into modern oncology workflows may open new opportunities for early detection, long-term surveillance, and precision treatment of BC.

Lina Jans, Jan Brynhildsen, Joar Hofgaard et al.

Abstract Background Users of intrauterine devices (IUDs) have been found to have a lower incidence of cervical cancer in meta-analyses, but these studies have not been able to examine the influence of IUD type. The aim of this study is to investigate the incidence of de novo high-risk human papillomavirus (HPV) infections in relation to the reported use of contraceptive methods, with special regard to different types of IUDs. Methods A sample of participants in the national screening program for cervical cancer (n = 11,702) with a negative HPV test in 2017–2018 were included. Their subsequent HPV test results in 2020–2023 were analyzed in relation to their reported contraceptive method. Results Participants who reported use of hormonal contraception had higher incidence of a positive HPV screening test (5.6%) compared with women with no reported contraception (4.2%) (OR 1.29; 95% CI 1.01–1.64). There was no significant difference in HPV incidence among women who reported use of hormonal IUD (HIUD) or copper-containing IUD (CU-IUD). Women who reported use of the same contraceptive method in both screening rounds showed no significant differences in HPV incidence, regardless of the contraceptive method they had used. Conclusion The incidence of de novo HPV infections is not significantly different in users of different types of IUD.

Yuming Zheng, Fen Xue, Dan Ou et al.

Abstract Purpose This study aims to evaluate the efficacy and toxicity of the two induction chemotherapy (IC) regimens (TPF: docetaxel, cisplatin and 5-fluorouracil, and PF: cisplatin and 5-fluorouracil) combined with radiotherapy in young and middle aged patients with locoregionally advanced nasopharyngeal carcinoma (LA-NPC). Methods A retrospective analysis was conducted on 329 cases with stage III-IVA nasopharyngeal carcinoma from September 2005 to February 2017. Of the 329 cases, 253 cases underwent TPF (docetaxel: 60 mg/m2 on day 1, cisplatin: 25 mg/m2 on days 1–3, 5-fluorouracil: 500 mg/m2 on days 1–5, intravenous 120-h infusion), while 76 cases received the PF regimen (cisplatin: 25 mg/m2 on days 1–3, 5-fluorouracil: 500 mg/m2 on days 1–5, intravenous 120-h infusion) every 3 weeks. Radiotherapy was administered after IC with or without concurrent chemotherapy. The survival rates were assessed by Kaplan–Meier analysis, and the survival curves were compared using a log‑rank test. Results The 5-year and 8-year overall survival (OS) rates of the PF group and TPF group were 80.1% and 72.1%, 87.3% and 78.4% respectively (p = 0.405). There were no statistical differences in regional recurrence-free survival (RRFS) and distant metastasis-free survival (DMFS) rates between PF and TPF groups(p = 0.585 and 0.500, respectively).The 5-year and 8-year estimated local recurrence free survival (LRFS) rates for patients in PF and TPF group were 91.1% and 78.0%, 96.2% and 93.7%, respectively (p = 0.026). Moreover, The OS, LRFS, RRFS and DMFS rates were comparable between the non CCRT or CCRT subgroup (p = 0.542, 0.319, 0.070, 0.986, respectively). Compared with PF group, the TPF group significantly increased the occurrence of grade 3 or 4 neutropenia and leukopenia (p < 0.001). Conclusion PF and TPF followed by radiotherapy with or without concurrent chemotherapy performed encouraging anti-tumor effects in LA-NPC, there was no statistical significance in 5-year and 8-year OS, RRFS, and DMFS rates between two chemotherapy regimens. Compared with PF, TPF induction chemotherapy achieved more satisfactory LRFS rate in LA-NPC with acceptable toxicity.

Richell Booijink, Leon W. M. M. Terstappen, Eshwari Dathathri et al.

In prostate cancer (PCa), cancer‐associated fibroblasts (CAFs) promote tumor progression, drug resistance, and metastasis. Although circulating tumor cells are studied as prognostic and diagnostic markers, little is known about other circulating cells and their association with PCa metastasis. Here, we explored the presence of circulating CAFs (cCAFs) in metastatic castration‐naïve prostate cancer (mCNPC) patients. cCAFs were stained with fibroblast activation protein (FAP), epithelial cell adhesion molecule (EpCAM), and receptor‐type tyrosine‐protein phosphatase C (CD45), then FAP+EpCAM− cCAFs were enumerated and sorted using fluorescence‐activated cell sorting. FAP+EpCAM− cCAFs ranged from 60 to 776 (389 mean ± 229 SD) per 2 × 108 mononuclear cells, whereas, in healthy donors, FAP+ EpCAM− cCAFs ranged from 0 to 71 (28 mean ± 22 SD). The mCNPC‐derived cCAFs showed positivity for vimentin and intracellular collagen‐I. They were viable and functional after sorting, as confirmed by single‐cell collagen‐I secretion after 48 h of culturing. Two cCAF subpopulations, FAP+CD45− and FAP+CD45+, were identified, both expressing collagen‐I and vimentin, but with distinctly different morphologies. Collectively, this study demonstrates the presence of functional and viable circulating CAFs in mCNPC patients, suggesting the role of these cells in prostate cancer.

Arhit Chakrabarti, Yang Ni, Debdeep Pati et al.

We consider the problem of clustering grouped data for which the observations may include group-specific variables in addition to the variables that are shared across groups. This type of data is common in cancer genomics where the molecular information is usually accompanied by cancer-specific clinical information. Existing grouped clustering methods only consider the shared variables, thereby ignoring valuable information from the cancer-specific variables. To allow for these cancer-specific variables to aid in the clustering, we propose a novel Bayesian nonparametric approach, termed global-local (GLocal) Dirichlet process, that models the ``global-local'' structure of the observations across groups. We characterize the GLocal Dirichlet process using the stick-breaking representation and the representation as a limit of a finite mixture model, which leads to an efficient posterior inference algorithm. We illustrate our model with extensive simulations and a real pan-gastrointestinal cancer dataset. The cancer-specific clinical variables included carcinoembryonic antigen level, patients' body mass index, and the number of cigarettes smoked per day. These important clinical variables refine the clusters of gene expression data and allow us to identify finer sub-clusters, which is not possible in their absence. This refinement aids in the better understanding of tumor progression and heterogeneity. Moreover, our proposed method is applicable beyond the field of cancer genomics to a general grouped clustering framework in the presence of group-specific idiosyncratic variables.

Emmanuella Avwerosuoghene Oghenekaro

Early cancer detection remains one of the most critical challenges in modern healthcare, where delayed diagnosis significantly reduces survival outcomes. Recent advancements in artificial intelligence, particularly deep learning, have enabled transformative progress in medical imaging analysis. Deep learning-based computer vision models, such as convolutional neural networks (CNNs), transformers, and hybrid attention architectures, can automatically extract complex spatial, morphological, and temporal patterns from multimodal imaging data including MRI, CT, PET, mammography, histopathology, and ultrasound. These models surpass traditional radiological assessment by identifying subtle tissue abnormalities and tumor microenvironment variations invisible to the human eye. At a broader scale, the integration of multimodal imaging with radiogenomics linking quantitative imaging features with genomics, transcriptomics, and epigenetic biomarkers has introduced a new paradigm for personalized oncology. This radiogenomic fusion allows the prediction of tumor genotype, immune response, molecular subtypes, and treatment resistance without invasive biopsies.

Jason C. Hsu

In assessing Overall Survival (OS) in oncology studies, it is essential for the efficacy measure to be Logic-respecting, for otherwise patients may be incorrectly targeted. This paper explains, while Time Ratio (TR) is Logic-respecting, Hazard Ratio (HR) is not Logic-respecting. With Time Ratio (TR) being recommended, a smooth transitioning strategy is suggested. The conclusion states: Logicality requires, and Subgroup Mixable Estimation (SME) delivers, an efficacy assessment for the overall population within the range of minimum and maximum efficacy in the subgroups, no matter how outcome is measured, whichever logic-respecting efficacy measure is chosen, the same efficacy assessment regardless of how subgroups are stratified.

Lingjun Xiao, Louqian Zhang, Ciliang Guo et al.

Abstract Phagocytosis, a vital defense mechanism, involves the recognition and elimination of foreign substances by cells. Phagocytes, such as neutrophils and macrophages, rapidly respond to invaders; macrophages are especially important in later stages of the immune response. They detect “find me” signals to locate apoptotic cells and migrate toward them. Apoptotic cells then send “eat me” signals that are recognized by phagocytes via specific receptors. “Find me” and “eat me” signals can be strategically harnessed to modulate antitumor immunity in support of cancer therapy. These signals, such as calreticulin and phosphatidylserine, mediate potent pro‐phagocytic effects, thereby promoting the engulfment of dying cells or their remnants by macrophages, neutrophils, and dendritic cells and inducing tumor cell death. This review summarizes the phagocytic “find me” and “eat me” signals, including their concepts, signaling mechanisms, involved ligands, and functions. Furthermore, we delineate the relationships between “find me” and “eat me” signaling molecules and tumors, especially the roles of these molecules in tumor initiation, progression, diagnosis, and patient prognosis. The interplay of these signals with tumor biology is elucidated, and specific approaches to modulate “find me” and “eat me” signals and enhance antitumor immunity are explored. Additionally, novel therapeutic strategies that combine “find me” and “eat me” signals to better bridge innate and adaptive immunity in the treatment of cancer patients are discussed.

Bianca Medici, Eugenia Caffari, Massimiliano Salati et al.

Introduction: Spontaneous tumor regression is the volumetric reduction or complete disappearance of a primary tumor or metastatic sites (single or multiple) without the administration of treatments. This rare phenomenon occurs most commonly in certain types of neoplasms. Case Presentation: In this manuscript, we describe a spontaneous tumor regression in an adult patient followed at the Modena Cancer Center and affected by retroperitoneal inflammatory myofibroblastic tumor, an ultra-rare subtype of sarcoma. Finally, we will provide a concise review of the literature and try to explain the mechanisms underlying the tumor regression described in the clinical case. Conclusion: The etiopathogenetic mechanisms for spontaneous tumor regression are not yet fully understood and likely involve a complex interplay among immunological mechanisms, growth factors, cytokines, and hormonal factors.

Lidia Garrucho, Kaisar Kushibar, Claire-Anne Reidel et al.

Artificial Intelligence (AI) research in breast cancer Magnetic Resonance Imaging (MRI) faces challenges due to limited expert-labeled segmentations. To address this, we present a multicenter dataset of 1506 pre-treatment T1-weighted dynamic contrast-enhanced MRI cases, including expert annotations of primary tumors and non-mass-enhanced regions. The dataset integrates imaging data from four collections in The Cancer Imaging Archive (TCIA), where only 163 cases with expert segmentations were initially available. To facilitate the annotation process, a deep learning model was trained to produce preliminary segmentations for the remaining cases. These were subsequently corrected and verified by 16 breast cancer experts (averaging 9 years of experience), creating a fully annotated dataset. Additionally, the dataset includes 49 harmonized clinical and demographic variables, as well as pre-trained weights for a baseline nnU-Net model trained on the annotated data. This resource addresses a critical gap in publicly available breast cancer datasets, enabling the development, validation, and benchmarking of advanced deep learning models, thus driving progress in breast cancer diagnostics, treatment response prediction, and personalized care.

S. Ruff, D. Diaz, K. Pitter et al.

A 63‐year‐old woman who was a former smoker with a past medical history of hypertension and gastroesophageal reflux disease initially presented with upper abdominal pain. Her family history was notable for breast cancer in her mother, lung cancer in her father, and renal cell carcinoma in her sister. An ultrasound showed a heterogenous mass in the left lobe of the liver measuring 8.7 � 7.0 � 5.1 cm that was abutting the common bile duct and concerning for a neoplasm (Figure 1A). On laboratory testing, her alpha fetoprotein (AFP) was elevated (15.7 ng/mL), carbohydrate antigen 19‐9 (CA 19‐9) was normal (<15 U/ml), and carcinoembryonic antigen (CEA) was slightly elevated (0.6 ng/ml). She underwent an ultrasound‐guided biopsy that demonstrated cytokeratin 7 (CK7)‐positive, poorly differentiated adenocarcinoma with nonmucinous gland formation and papillary architecture within sclerotic stroma. Given that the biopsy was positive for CK7 with negative hepatocellular (hepatocyte‐specific antigen, arginase, glypican), CDX2, TTF1, and synaptophysin markers, the mass was diagnosed as an intrahepatic cholangiocarcinoma (iCCA). A computed tomography (CT) scan of the chest, abdomen, and pelvis did not show any extrahepatic metastatic disease but did show a central left hepatic lobe mass in segment 4a/4b that measured 7.7 � 6.7 cm with calcifications suggestive of iCCA (Figure 1B,C). A CT scan also revealed potential tumor thrombus within the middle hepatic vein and distal left portal vein branches, extrahepatic (periportal, gastrohepatic, peripancreatic, portacaval) lymphadenopathy, left intrahepatic biliary ductal dilation, and common bile duct dilation. The patient was started on gemcitabine, cisplatin, and nanoparticle albumin‐bound paclitaxel (nab‐paclitaxel). After 3 months of chemotherapy, the patient's AFP increased to 43.8 ng/ml and her CA 19‐9 increased to 21.4 U/ml. On repeat CT scan, the size of the tumor was stable, but there was suspected intraductal extension toward the central inferior aspect of segment 4b. Given her suboptimal response to chemotherapy, radiation oncology was consulted. Approximately 4 months after starting chemotherapy, the patient underwent yttrium‐90 radioembolization (Y90 RE) to the left hepatic hemiliver and subsequently was resumed on a gemcitabine, cisplatin, and nab‐paclitaxel regimen (Figure 2A). A CT scan 5 months after starting treatment and 1 month after Y90 RE demonstrated a stable left hepatic lobe mass with interval necrosis. However, this effect was mostly seen in the tumor in the left lobe of the liver, whereas there was still some residual arterial enhancement along the right side of the mass because where the tumor extended into the right lobe was not treated given concern of toxicity to the remaining liver. After nine cycles of chemotherapy and the Y90 RE treatment, re‐staging CT scans did not demonstrate any metastatic disease, and the tumor in the left lobe of the liver had a seemingly good response to the Y90 RE (Figure 2B). In addition, the periportal, portacaval, and gastrohepatic lymphadenopathy had decreased in size, and there was no new or progressive adenopathy. At this point, the patient was taken to the operating room, and an extended left hepatectomy, cholecystectomy, and extensive lymphadenectomy, including skeletonizing of the hilum, left hepatic artery, bile ducts, and common hepatic artery, was performed. On postoperative day 5, the patient was tachycardic, a CT scan showed a

Natascha Roehlen, M. Muller, Z. Nehme et al.

BACKGROUND AND AIMS Hepatocellular carcinoma (HCC) is a leading cause of cancer-related death worldwide. Despite new treatment approvals, treatment response and prognosis of patients with advanced HCC remain poor. Claudin-1 (CLDN1) is a membrane protein expressed not only at tight junctions but also non-junctionally such as the basolateral membrane of the human hepatocyte. While CLDN1 within tight junctions is well characterized, the role of non-junctional CLDN1 and its role as a therapeutic target in HCC remains unexplored. METHODS Using humanized monoclonal antibodies (mAbs) targeting specifically the extracellular loop of human non-junctional CLDN1 and a large series of patient-derived cell-based and animal model systems we aimed to investigate the role of CLDN1 as a therapeutic target for HCC. RESULTS Targeting non-junctional CLDN1 markedly suppressed tumor growth and invasion in cell line-based models of HCC and patient-derived 3D ex vivo models. Moreover, the robust effect on tumor growth was confirmed in vivo in a large series of cell line-derived xenograft (CDX) and patient-derived xenograft (PDX) mouse models. Mechanistic studies including single-cell RNA sequencing of multicellular patient HCC tumorspheres suggested that CLDN1 regulates tumor stemness, metabolism, oncogenic signaling and perturbs the tumor immune microenvironment. CONCLUSIONS Our results provide the rationale for targeting CLDN1 in HCC and pave the way to novel therapeutic interventions with CLDN1 mAbs aimed at improving the limited efficacy of current therapies. IMPACT AND IMPLICATIONS Hepatocellular carcinoma (HCC) is a cancer with high mortality and unsatisfactory treatment options. Here we identified the cell surface protein Claudin-1 as a target for treatment of advanced HCC. Monoclonal antibodies targeting Claudin-1 inhibit tumor growth in patient-derived ex vivo and in vivo models by modulating signaling, cell stemness and the tumor immune microenvironment. Given the differentiated mechanism of action, the identification of Claudin-1 as a novel therapeutic target for HCC provides an opportunity to break the plateau of limited treatment response. These results of this preclinical study pave the way for the clinical development of Claudin-1 specific antibodies for treatment of HCC in patients. It is therefore of key impact for physicians, scientists and drug developers in the field of liver cancer and GI oncology.

Dimitra Galiti, Helena Linardou, Sofia Agelaki et al.

Introduction: The increasing burden of cancer, the development of novel therapies, and the COVID-19 pandemic have made cancer care more complex. Digital innovation was then pushed toward developing platforms to facilitate access to cancer care. Age, education, and other disparities were, however, shown to limit the use of the digital health innovation. The aim of this early-stage feasibility study was to assess whether Greek cancer patients would register at CureCancer and self-report their demographics, disease and therapy characteristics, and socioeconomic issues. The study was organized by the Hellenic Society of Medical Oncology. Methods: Patients from nine cancer centers were invited to register on the CureCancer platform and complete an anonymous questionnaire on demographics, disease and therapy characteristics, and socioeconomic issues. Patients were also encouraged to upload, in a secure area for them, their medical files and share them with their physicians. They were then asked to comment on their experience of registration and how easy it was to upload their medical files. Results: Of the 159 patients enrolled, 144 (90.56%) registered, and 114 of those (79.16%) completed the questionnaire, suggesting that the study is feasible. Users’ median age was 54.5 years, and 86.8% of them were university and high school graduates. Most patients (79.8%) reported their specific type of cancer diagnosis, and all reported their therapy characteristics. Breast and lung cancers were the most common. A total of 87 patients (76.3%) reported being on active cancer therapy, 46 (40.4%) had metastatic disease, and 51 (44.7%) received supportive care medications. Eighty-one (71.05%) patients received prior cancer therapies, and twenty-seven recalled prior supportive care medications. All patients reported visiting non-oncology Health Care Professionals during the study. Nineteen of 72 (26.39%) patients who worked prior to cancer diagnosis changed work status; 49 (42.98) patients had children under 24 years; and 16 (14%) patients lived alone. Nine (7.9%) patients were members of patient associations. Registration was “much/very much” easy for 98 (86.0%) patients, while 67 (58.8%) had difficulties uploading their files. Patients commented on the well-organized data access, improved communication, feeling safe, medication adherence, interventions from a distance, and saving time and money. Over 80% of patients “preferred the digital way”. Discussion: A total of 114 patients succeeded in registering on the digital platform and reporting their demographics, disease and therapy characteristics, and socioeconomic issues. Age and educational disparities were disclosed and highlighted the need for educational programs to help older people and people of lower education use digital innovation. Health care policy measures would support patients’ financial burden associated with work changes, living alone, and children under 24 years old at school or college. Policy actions would motivate patients to increase their participation in patient associations. According to the evidence DEFINED framework, the number of patients, and the focus on enrollment, engagement, and user experience, the study fulfills actionability level criterion 1.

An-An Xu, Jing-Jing Miao, Lin Wang et al.

Abstract Background To analysis the clinical outcomes of concurrent chemoradiotherapy (CCRT) alone based on 10-year results for loco-regionally advanced nasopharyngeal carcinoma (LANPC), so as to provide evidence for individualized treatment strategy and designing appropriate clinical trial for different risk LANPC patients. Methods Consecutive patients with stage III-IVa (AJCC/UICC 8th) were enrolled in this study. All patients received radical intensity-modulated radiotherapy (IMRT) and concurrent cisplatin chemotherapy (CDDP). The hazard ratios (HRs) of death risk in patients with T3N0 was used as baseline, relative HRs were calculated by a Cox proportional hazard model to classify different death risk patients. Survival curves for the time-to-event endpoints were analyzed by the Kaplan–Meier method and compared using the log-rank test. All statistical tests were conducted at a two-sided level of significance of 0.05. Results A total of 456 eligible patients were included. With 12-year median follow-up, 10-year overall survival (OS) was 76%. 10-year loco-regionally failure-free survival (LR-FFS), distant failure-free survival (D-FFS) and failure-free survival (FFS) were 72%, 73% and 70%, respectively. Based on the relative hazard ratios (HRs) of death risk, LANPC patients were classified into 3 subgroups, low-risk group (T1-2N2 and T3N0-1) contained 244 patients with HR < 2; medium-risk group (T3N2 and T4N0-1) contained 140 patients with HR of 2 – 5; high-risk group (T4N2 and T1-4N3) contained 72 patients with HR > 5. The 10-year OS for patients in low-, medium-, and high-risk group were 86%, 71% and 52%, respectively. Significantly differences of OS rates were found between each of the two groups (low-risk group vs. medium-risk group, P < 0.001; low-risk group vs. high-risk group, P < 0.001; and medium-risk group vs. high-risk group, P = 0.002, respectively). Grade 3–4 late toxicities included deafness/otitis (9%), xerostomia (4%), temporal lobe injury (5%), cranial neuropathy (4%), peripheral neuropathy (2%), soft tissue damage (2%) and trismus (1%). Conclusions Our classification criteria demonstrated that significant heterogeneity in death risk among TN substages for LANPC patients. IMRT plus CDDP alone maybe suitable for low-risk LANPC (T1-2N2 or T3N0-1), but not for medium- and high-risk patients. These prognostic groupings provide a practicable anatomic foundation to guide individualized treatment and select optimal targeting in the future clinical trials.

Hui Li, Chao Yang, Zhonghong Liao et al.

ObjectiveTo compare our initial perioperative and postoperative outcomes of the modified anterior approach (MA) with Retzius space preservation robot-assisted radical prostatectomy (RARP) with the standard anterior approach (SA) RARP.Materials and methodsA retrospective analysis was performed on 116 patients with RARP completed by the same surgeon between September 2019 and March 2022. They were divided into SA-RARP group (77 cases) and MA-RARP group (39 cases). Propensity score matching was performed using eight preoperative variables, including age, BMI, preoperative PSA, biopsy Gleason score, prostate volume, D’Amico risk classification, SHIM, and clinical T stage. Functional outcome was assessed by urine pad count and SHIM after surgery, and oncological outcome was assessed by statistics of postoperative pathological findings as well as follow-up postoperative PSA. The median follow-up was 13 months and 17 months for MA-RARP and SA-RARP groups respectively.ResultsPropensity score matching was performed 1:1, and baseline data were comparable between the two groups after matching. Comparison of postoperative data: MA-RARP group had less mean EBL than SA-RARP group (200 vs 150 ml, p = 0.033). PSM did not differ between groups (p = 1). In terms of urinary control recovery, the MA-RARP group showed significant advantages in urinary control recovery at 24 h, 2 weeks, 1 month and 3 months after catheter removal, respectively (48.6% vs 5.7%, p &lt; 0.001; 80% vs 22.9%, p &lt; 0.001; 94.3% vs 51.4%, p &lt; 0.001; 100% vs 74.3%, p = 0.002). This advantage gradually disappeared 6 months or more after surgery. The median time to recovery of sexual function was shorter in the MA-RARP group (165 vs 255 d, p = 0.001).ConclusionMA-RARP is safe and reliable, and can achieve better early urinary control function and sexual function recovery while achieving the primary tumor control goal.

Alexandra Erdmann, Jan Beyersmann, Kaspar Rufibach

When planning an oncology clinical trial, the usual approach is to assume proportional hazards and even an exponential distribution for time-to-event endpoints. Often, besides the gold-standard endpoint overall survival (OS), progression-free survival (PFS) is considered as a second confirmatory endpoint. We use a survival multistate model to jointly model these two endpoints and find that neither exponential distribution nor proportional hazards will typically hold for both endpoints simultaneously. The multistate model provides a stochastic process approach to model the dependency of such endpoints neither requiring latent failure times nor explicit dependency modelling such as copulae. We use the multistate model framework to simulate clinical trials with endpoints OS and PFS and show how design planning questions can be answered using this approach. In particular, non-proportional hazards for at least one of the endpoints are naturally modelled as well as their dependency to improve planning. We consider an oncology trial on non-small-cell lung cancer as a motivating example from which we derive relevant trial design questions. We then illustrate how clinical trial design can be based on simulations from a multistate model. Key applications are co-primary endpoints and group-sequential designs. Simulations for these applications show that the standard simplifying approach may very well lead to underpowered or overpowered clinical trials. Our approach is quite general and can be extended to more complex trial designs, further endpoints, and other therapeutic areas. An R package is available on CRAN.

Bliss Singhal, Fnu Pooja

Machine learning (ML) is a branch of Artificial Intelligence (AI) where computers analyze data and find patterns in the data. The study focuses on the detection of metastatic cancer using ML. Metastatic cancer is the point where the cancer has spread to other parts of the body and is the cause of approximately 90% of cancer related deaths. Normally, pathologists spend hours each day to manually classify whether tumors are benign or malignant. This tedious task contributes to mislabeling metastasis being over 60% of time and emphasizes the importance to be aware of human error, and other inefficiencies. ML is a good candidate to improve the correct identification of metastatic cancer saving thousands of lives and can also improve the speed and efficiency of the process thereby taking less resources and time. So far, deep learning methodology of AI has been used in the research to detect cancer. This study is a novel approach to determine the potential of using preprocessing algorithms combined with classification algorithms in detecting metastatic cancer. The study used two preprocessing algorithms: principal component analysis (PCA) and the genetic algorithm to reduce the dimensionality of the dataset, and then used three classification algorithms: logistic regression, decision tree classifier, and k-nearest neighbors to detect metastatic cancer in the pathology scans. The highest accuracy of 71.14% was produced by the ML pipeline comprising of PCA, the genetic algorithm, and the k-nearest neighbors algorithm, suggesting that preprocessing and classification algorithms have great potential for detecting metastatic cancer.

Sho Komukai, Satoshi Hattori, Bernard Rachet

In epidemiology research with cancer registry data, it is often of primary interest to make inference on cancer death, not overall survival. Since cause of death is not easy to collect or is not necessarily reliable in cancer registries, some special methodologies have been introduced and widely used by using the concepts of the relative survival ratio and the net survival. In making inference of those measures, external life tables of the general population are utilized to adjust the impact of non-cancer death on overall survival. The validity of this adjustment relies on the assumption that mortality in the external life table approximates non-cancer mortality of cancer patients. However, the population used to calculate a life table may include cancer death and cancer patients. Sensitivity analysis proposed by Talbäck and Dickman to address it requires additional information which is often not easily available. We propose a method to make inference on the net survival accounting for potential presence of cancer patients and cancer death in the life table for the general population. The idea of adjustment is to consider correspondence of cancer mortality in the life table and that in the cancer registry. We realize a novel method to adjust cancer mortality in the cancer registry without any additional information to the standard analyses of cancer registries. Our simulation study revealed that the proposed method successfully removed the bias. We illustrate the proposed method with the cancer registry data in England.

Omar S. Ads, Mayar M. Alfares, Mohammed A. -M. Salem

Brain tumors are one of the life-threatening forms of cancer. Previous studies have classified brain tumors using deep neural networks. In this paper, we perform the later task using a collaborative deep learning technique, more specifically split learning. Split learning allows collaborative learning via neural networks splitting into two (or more) parts, a client-side network and a server-side network. The client-side is trained to a certain layer called the cut layer. Then, the rest of the training is resumed on the server-side network. Vertical distribution, a method for distributing data among organizations, was implemented where several hospitals hold different attributes of information for the same set of patients. To the best of our knowledge this paper will be the first paper to implement both split learning and vertical distribution for brain tumor classification. Using both techniques, we were able to achieve train and test accuracy greater than 90\% and 70\%, respectively.

R. Short, M. Greenwade, A. Bonebrake