WITHDRAWN: Neoadjuvant therapy leads to objective response in intrahepatic cholangiocarcinoma.

The Publisher regrets that this article is an accidental duplication of an article that has already been published, https://doi.org/10.1016/j.hpb.2024.04.011. The duplicate article has therefore been withdrawn. The full Elsevier Policy on Article Withdrawal can be found at https://www.elsevier.com/about/policies/article-withdrawal.

Neoadjuvant therapy leads to objective response in intrahepatic cholangiocarcinoma.

BACKGROUND: Intrahepatic cholangiocarcinoma (iCCA) is the second most common hepatic malignancy and has a poor prognosis. Surgical resection is the standard of care for patients with resectable disease, representing 30-40% of cases. Increasingly, neoadjuvant systemic therapy is being utilized in patients due to high-risk anatomic or biologic considerations. However, data on the clinical effect of this approach are limited. We performed a cohort study to evaluate the effect of neoadjuvant therapy in patients with oncologically high-risk iCCA. METHODS: iCCA patients (n = 181) between the years 2014-2020 were reviewed for clinical, histopathologic, treatment, and outcome-related data. Tumor regression grade was scored per CAP criteria for gastrointestinal carcinomas. RESULTS: 47 iCCA patients received neoadjuvant therapy and 72 did not. Neoadjuvant treatment led to objective response and tumor regression by CAP score. After adjustment for age, clinical stage, and tumor size, the outcomes of patients who had neoadjuvant therapy followed by surgery were not significantly different from those patients who had surgery first. DISCUSSION: In conclusion, neoadjuvant therapy in iCCA facilitated surgical care. The progression-free and overall survival for surgical patients with and without neoadjuvant therapy were not significantly different suggesting this approach needs further exploration as an effective treatment paradigm.

Hepatocellular carcinoma downstaging for liver transplantation in the era of systemic combined therapy with anti-VEGF/TKI and immunotherapy.

Hepatocellular carcinoma remains a global health challenge affecting close to 1 million cases yearly. Liver transplantation provides the best long-term outcomes for those meeting strict criteria. Efforts have been made to expand these criteria, whereas others have attempted downstaging approaches. Although locoregional approaches to downstaging are appealing and have demonstrated efficacy, limitations and challenges exists including poor imaging modality to assess response and appropriate endpoints along the process. Recent advances in systemic treatments including immune checkpoint inhibitors alone or in combination with tyrosine kinase inhibitors have prompted the discussion regarding their role for downstaging disease prior to transplantation. Here, we provide a review of prior locoregional approaches for downstaging, new systemic agents and their role for downstaging, and finally, key and critical considerations of the assessment, endpoints, and optimal designs in clinical trials to address this key question.

Transfer function approach to understanding periodic forcing of signal transduction networks.

Signal transduction networks are responsible for transferring biochemical signals from the extracellular to the intracellular environment. Understanding the dynamics of these networks helps understand their biological processes. Signals are often delivered in pulses and oscillations. Therefore, understanding the dynamics of these networks under pulsatile and periodic stimuli is useful. One tool to do this is the transfer function. This tutorial outlines the basic theory behind the transfer function approach and walks through some examples of simple signal transduction networks.

Social determinants of health: a need for better data capture in Asian American patients with hepatocellular cancer.

BACKGROUND: Social determinants of health lead to better cancer care. This multi-site, single-institution study sought to capture data on social determinants of health data in Asian Americans with hepatocellular carcinoma; this group constitutes 60% of patients with this malignancy and are often undertreated or not treated at all. METHODS: This study took advantage of an institutional initiative designed to capture and integrate social determinants of health data into the electronic medical record for all patients. Medical records of Asian Americans with hepatocellular cancer were reviewed to acquire data on housing instability, lack of transportation, financial concerns, and social isolation; a score of 1 indicated poor social determinants of health. RESULTS: Of 112 adult Asian American patients with hepatocellular cancer, 22 (20%) were Southeast Asian, and 74 (67%) described English proficiency/preference. Total noncompletion per domain (no question answered within that domain) was observed in 90 patients (80%) for housing instability; 90 (80%) for lack of transportation; 92 (82%) for financial hardship; and 90 (80%) for social isolation. A score of 1 (highest risk) was observed in 1 patient (0.9%) for housing instability; 1 (0.9%) lack of transportation; no patient for financial hardship; and 1 (0.9%) for social isolation. Of note, institution-wide benchmark total noncompletion rates were 0.3%, 0.3%, 47%, and 39% for these respective domains. CONCLUSION: High total noncompletion rates make social determinants of health data challenging to interpret and underscore the need for evidence-based guidelines on how best to capture such data in underserved patients.

Gastrointestinal perforation after bevacizumab: a multi-site, single-institution study with a focus on survival.

BACKGROUND: Bevacizumab-induced gastrointestinal perforation is a rare but potentially devastating adverse event that has generated limited data on overall survival. Yet, such survival data are critical in guiding management. METHODS: This multi-site, single-institution retrospective study focused on all cancer patients who had received bevacizumab and who had suffered a well-documented gastrointestinal perforation from January 1, 2004 through January 20, 2022.The main goal was to report survival outcomes; Kaplan Meier curves and Cox survival models were used for this purpose. RESULTS: Eighty-nine patients are included in this report with a median age of 62 years (range 26-85). Colorectal cancer was the most common malignancy (n = 42). Thirty-nine patients underwent surgery for the perforation. Seventy-eight were deceased at the time of reporting with an overall median survival of all patients of 2.7 months (range 0-45 months), and 32 (36%) died within 30 days of perforation. In univariable survival analyses, no statistically significant associations were observed for age, gender, corticosteroid use, and time since last bevacizumab dose. However, surgically treated patients manifested a better survival (hazard ratio (HR) 0.49 (95% CI 0.31-0.78); p = 0.003). In multivariable analyses, surgery continued to be associated with improved survival (HR 0.47 (95% CI 0.29-0.74); p = 0.002), and corticosteroid use was associated with worse survival (HR 1.75 (95% CI 1.02-2.99); p = 0.04). CONCLUSION: Although gastrointestinal perforation after bevacizumab should be managed on a case-by-case basis, these descriptive survival data can help inform patients, their families, and healthcare providers as challenging management decisions arise.

The role of microbiome in pancreatic cancer.

Recent studies of the human microbiome have offered new insights into how the microbiome can impact cancer development and treatment. Specifically, in pancreatic ductal adenocarcinoma (PDAC), the microbiota has been shown to modulate PDAC risk, contribute to tumorigenesis, impact the tumor microenvironment, and alter treatment response. These findings provide rationale for further investigations into leveraging the microbiome to develop new strategies to diagnose and treat PDAC patients. There is growing evidence that microbiome analyses have the potential to become easily performed, non-invasive diagnostic, prognostic, and predictive biomarkers in pancreatic cancer. More excitingly, there is now emerging interest in developing interventions based on the modulation of microbiota. Fecal microbiota transplantation, probiotics, dietary changes, and antibiotics are all potential strategies to augment the efficacy of current therapeutics and reduce toxicities. While there are still challenges to overcome, this is a rapidly growing field that holds promise for translation into clinical practice and provides a new approach to improving patient outcomes.

A pilot study of Pan-FGFR inhibitor ponatinib in patients with FGFR-altered advanced cholangiocarcinoma.

Background Biliary tract cancers (BTC) are rare, chemo resistant and are associated with a poor prognosis. Preclinical and early clinical work had demonstrated interesting anti-tumor activity from targeting fibroblast growth factor receptor (FGFR) pathway. We hypothesized that ponatinib, a multi-targeted tyrosine kinase inhibitor with activity against FGFR, would be active in BTC patients with FGFR alterations. Methods This was a multi-center, single institution pilot study of ponatinib in patients with advanced, refractory BTC with FGFR alterations. The primary end point was overall response rate, with secondary points of overall survival (OS), progression-free survival (PFS) and Health Related Quality of Life (HRQoL) assessment. Results Twelve patients were enrolled prior to early termination of the trial. Partial responses were observed in 1 from 12 patients. Median PFS was 2.4 months and median OS was 15.7 months. All observed toxicities were manageable and reversible. Toxicities were mild, with lymphopenia (75%), rash (63%) and fatigue (50%) being the most frequent. No significant detriment in global QoL was observed. Conclusions Ponatinib as a single agent in FGFR altered BTC is tolerable with limited clinical activity. This is the first report of prospective assessment of FGFR inhibition in BTC using ponatinib, and the first study to report its effect on HRQoL. Further development of ponatinib will involve correlative studies to better refine patient selection, focus on combinations with other molecular targeted agents, conventional cytotoxic chemotherapy, and studies to better understand mechanisms of treatment resistance.

FDG-PET Predicts Neoadjuvant Therapy Response and Survival in Borderline Resectable/Locally Advanced Pancreatic Adenocarcinoma.

BACKGROUND: Neoadjuvant therapy (NAT) is used in borderline resectable/locally advanced (BR/LA) pancreatic ductal adenocarcinoma (PDAC). Anatomic imaging (CT/MRI) poorly predicts response, and biochemical (CA 19-9) markers are not useful (nonsecretors/nonelevated) in many patients. Pathologic response highly predicts survival post-NAT, but is only known postoperatively. Because metabolic imaging (FDG-PET) reveals primary tumor viability, this study aimed to evaluate our experience with preoperative FDG-PET in patients with BR/LA PDAC in predicting NAT response and survival. METHODS: We reviewed all patients with resected BR/LA PDAC who underwent NAT with FDG-PET within 60 days of resection. Pre- and post-NAT metabolic (FDG-PET) and biochemical (CA 19-9) responses were dichotomized in addition to pathologic responses. We compared post-NAT metabolic and biochemical responses as preoperative predictors of pathologic responses and recurrence-free survival (RFS) and overall survival (OS). RESULTS: We identified 202 eligible patients. Post-NAT, 58% of patients had optimization of CA 19-9 levels. Major metabolic and pathologic responses were present in 51% and 38% of patients, respectively. Median RFS and OS times were 21 and 48.7 months, respectively. Metabolic response was superior to biochemical response in predicting pathologic response (area under the curve, 0.86 vs 0.75; P<.001). Metabolic response was the only univariate preoperative predictor of OS (odds ratio, 0.25; 95% CI, 0.13-0.40), and was highly correlated (P=.001) with pathologic response as opposed to biochemical response alone. After multivariate adjustment, metabolic response was the single largest independent preoperative predictor (P<.001) for pathologic response (odds ratio, 43.2; 95% CI, 16.9-153.2), RFS (hazard ratio, 0.37; 95% CI, 0.2-0.6), and OS (hazard ratio, 0.21; 95% CI, 0.1-0.4). CONCLUSIONS: Among patients with post-NAT resected BR/LA PDAC, FDG-PET highly predicts pathologic response and survival, superior to biochemical responses alone. Given the poor ability of anatomic imaging or biochemical markers to assess NAT responses in these patients, FDG-PET is a preoperative metric of NAT efficacy, thereby allowing potential therapeutic alterations and surgical treatment decisions. We suggest that FDG-PET should be an adjunct and recommended modality during the NAT phase of care for these patients.

Shifting Epidemiology of Hepatocellular Carcinoma in Far Eastern and Southeast Asian Patients: Explanations and Implications.

PURPOSE OF REVIEW: Liver cancer is a major health burden globally with high rates in the Far East and Southeast Asia. Hepatitis B virus and hepatitis C virus remain the major causes of hepatocellular carcinoma (HCC) worldwide; however, the epidemiology is changing. Here, we review the changing epidemiological changes among Far East and Southeast Asians and among Asian Americans in the USA. RECENT FINDINGS: HBV- and HCV-induced HCC is anticipated to decline in incidence. In contrast, nonalcoholic fatty liver disease-induced HCC will outrank viral hepatitis in the next decade. Alcoholic liver disease will remain a persistent risk factor for HCC. These epidemiological shifts might have implications for the care of patients with HCC as described in this review. Societal changes are impacting the global incidence of HCC and will likely also impact the care of patients with this malignancy.

Phase IB study of sorafenib and evofosfamide in patients with advanced hepatocellular and renal cell carcinomas (NCCTG N1135, Alliance).

Background Sorafenib (Sor) remains a first-line option for hepatocellular carcinoma (HCC) or refractory renal cell carcinomas (RCC). PLC/PRF/5 HCC model showed upregulation of hypoxia with enhanced efficacy when Sor is combined with hypoxia-activated prodrug evofosfamide (Evo). Methods This phase IB 3 + 3 design investigated 3 Evo dose levels (240, 340, 480 mg/m2 on days 8, 15, 22), combined with Sor 200 mg orally twice daily (po bid) on days 1-28 of a 28-day cycle. Primary objectives included determining maximum tolerated dose (MTD) and recommended phase II dose (RP2D) of Sor + Evo. Results Eighteen patients were enrolled (median age 62.5 years; 17 male /1 female; 12 HCC/6 RCC) across three dose levels (DL0: Sor 200 mg bid/Evo 240 mg/m2 [n = 6], DL1:Sor 200 mg bid/Evo 480 mg/m2 [n = 5], DL1a: Sor 200 mg bid/Evo 340 mg/m2 [n = 7]). Two dose-limiting toxicities (DLTs) were reported with Evo 480 mg/m2 (grade 3 mucositis, grade 4 hepatic failure). Grade 3 rash DLT was observed in one patient at Evo 240 mg/m2. No DLTs were observed at Evo 340 mg/m2. MTD and RP2D were established as Sor 200 mg/Evo 340 mg/m2 and Sor 200/Evo 240 mg/m2, respectively. The most common treatment-related adverse events included fatigue, hand-foot syndrome, hypertension, and nausea/vomiting. Two partial responses were observed, one each at DL0 and DL1a.; disease control rate was 55%. Conclusions RP2D was established as sorafenib 200 mg bid + Evo 240 mg/m2. While preliminary anti-tumor activity was observed, future development must account for advances in immunotherapy in HCC/RCC.

Self-Coexistence among IEEE 802.22 Networks: Distributed Allocation of Power and Channel.

Ensuring self-coexistence among IEEE 802.22 networks is a challenging problem owing to opportunistic access of incumbent-free radio resources by users in co-located networks. In this study, we propose a fully-distributed non-cooperative approach to ensure self-coexistence in downlink channels of IEEE 802.22 networks. We formulate the self-coexistence problem as a mixed-integer non-linear optimization problem for maximizing the network data rate, which is an NP-hard one. This work explores a sub-optimal solution by dividing the optimization problem into downlink channel allocation and power assignment sub-problems. Considering fairness, quality of service and minimum interference for customer-premises-equipment, we also develop a greedy algorithm for channel allocation and a non-cooperative game-theoretic framework for near-optimal power allocation. The base stations of networks are treated as players in a game, where they try to increase spectrum utilization by controlling power and reaching a Nash equilibrium point. We further develop a utility function for the game to increase the data rate by minimizing the transmission power and, subsequently, the interference from neighboring networks. A theoretical proof of the uniqueness and existence of the Nash equilibrium has been presented. Performance improvements in terms of data-rate with a degree of fairness compared to a cooperative branch-and-bound-based algorithm and a non-cooperative greedy approach have been shown through simulation studies.

Conversion of glycerol to hydrogen rich gas.

Presently there is a glut of glycerol as the by-product of biofuel production and it will grow as production increases. The conundrum is how we can consume this material and convert it into a more useful product. One potential route is to reform glycerol to hydrogen rich gas including synthesis gas (CO + H2) and hydrogen. However, there is recent literature on various reforming techniques which may have a bearing on the efficiency of such a process. Hence in this review reforming of glycerol at room temperature (normally photo-catalytic), catalysis at moderate and high temperature and a non-catalytic pyrolysis process are presented. The high temperature processes allow the generation of synthesis gas with the hydrogen to carbon monoxide ratios being suitable for synthesis of dimethyl ether, methanol and for the Fischer-Tropsch process using established catalysts. Efficient conversion of synthesis gas to hydrogen involves additional catalysts that assist the water gas shift reaction, or involves in situ capture of carbon dioxide and hydrogen. Reforming at reduced temperatures including photo-reforming offers the opportunity of producing synthesis gas or hydrogen using single catalysts. Together, these processes will assist in overcoming the worldwide glut of glycerol, increasing the competitiveness of the biofuel production and reducing our dependency on the fossil based, hydrogen rich gas.

Second-Line Treatment of Pancreatic Adenocarcinoma: Shedding Light on New Opportunities and Key Talking Points from Clinical Trials.

Despite improvements in overall cancer mortality, deaths related to pancreatic cancer continue to rise. Following first-line treatment, second-line options are significantly limited. Classically, first-line treatment consisted of either gemcitabine or 5-fluorouracil based systemic chemotherapy. Upon progression of disease or recurrence, subsequent second-line treatment is still gemcitabine or 5-fluorouracil based chemotherapy, depending on what was used in the first line and the timing of progression or recurrence. A better understanding of the molecular underpinnings of pancreatic adenocarcinoma (PDAC) has led to new treatment strategies including specifically targeting the desmoplastic stroma, cytokine signaling and actionable mutations. Furthermore, efforts are also directed to enhance the immunogenicity profile of PDAC's well-established immunologically "cold" tumor microenvironment. More recently, the outstanding response rates of chimeric antigen receptor T (CAR-T) cells in hematologic malignancies, have led to clinical trials to evaluate the treatment modality in PDAC. In this review, we summarize recently presented clinical trials for metastatic pancreatic adenocarcinoma with novel treatment approaches in the second line and beyond.

Self-Organizing Democratized Learning: Toward Large-Scale Distributed Learning Systems.

Emerging cross-device artificial intelligence (AI) applications require a transition from conventional centralized learning systems toward large-scale distributed AI systems that can collaboratively perform complex learning tasks. In this regard, democratized learning (Dem-AI) lays out a holistic philosophy with underlying principles for building large-scale distributed and democratized machine learning systems. The outlined principles are meant to study a generalization in distributed learning systems that go beyond existing mechanisms such as federated learning (FL). Moreover, such learning systems rely on hierarchical self-organization of well-connected distributed learning agents who have limited and highly personalized data and can evolve and regulate themselves based on the underlying duality of specialized and generalized processes. Inspired by Dem-AI philosophy, a novel distributed learning approach is proposed in this article. The approach consists of a self-organizing hierarchical structuring mechanism based on agglomerative clustering, hierarchical generalization, and corresponding learning mechanism. Subsequently, hierarchical generalized learning problems in recursive forms are formulated and shown to be approximately solved using the solutions of distributed personalized learning problems and hierarchical update mechanisms. To that end, a distributed learning algorithm, namely DemLearn, is proposed. Extensive experiments on benchmark MNIST, Fashion-MNIST, FE-MNIST, and CIFAR-10 datasets show that the proposed algorithm demonstrates better results in the generalization performance of learning models in agents compared to the conventional FL algorithms. The detailed analysis provides useful observations to further handle both the generalization and specialization performance of the learning models in Dem-AI systems.

"It's Like You Stuck a Pin in It:" African American/Black Patients Describe Cutaneous Toxicity From Epidermal Growth Factor Receptor Inhibitors.

Background: Epidermal growth factor receptor (EGFR) inhibitors cause cutaneous toxicity in over 90% of patients. Conceivably, healthcare providers could overlook such toxicity in African American/Black patients because of a darker complexion. This qualitative study sought to learn about such cutaneous signs and symptoms and, if present, to report them in patients' own words. Methods: Any patient who self-identified as African American/Black and who had been prescribed an EGFR inhibitor was eligible. The current report focuses on patients' responses to the following question, "What have you noticed since starting your cancer treatment (the EGFR inhibitor), any particular symptoms or reactions, positive or negative?" All interview data were audio-recorded, transcribed, and then independently coded and analyzed by two investigators. Results: Fifteen patients are the focus of this report, and all described cutaneous toxicity. Patients appeared troubled by the cosmetic aspect of these drug-induced skin changes, including their acneiform appearance, describing "little pimples with little, little pus in it." Notable were comments on hyperpigmentation, "I'm a black person but…. became darker." Furthermore, patients experienced physical symptoms: "it itches;" "it's like you stuck a pin in it;" "stinging;" and "burning;". Conclusion: Although cutaneous toxicity from EGFR inhibitors might be more difficult to visualize among darkly complected patients, the graphic descriptions offered in this qualitative study underscore the need for clinicians to heighten their awareness of such toxicity in African American/Black patients.

Social Determinants of Health: A Need for Evidence-Based Guidelines on How to Capture Data on Underserved Patients.

Background . Social determinants of health lead to better cancer care. This multi-site, single-institution study sought to capture data on social determinants of health data in Asian Americans with hepatocellular carcinoma; this group constitutes 60% of patients with this malignancy and are often undertreated or not treated at all. Methods . This study took advantage of an institutional initiative designed to capture and integrate social determinants of health data into the electronic medical record for all patients. Medical records of Asian Americans with hepatocellular cancer were reviewed to acquire data on housing instability, lack of transportation, financial concerns, and social isolation; a score of 1 indicated poor social determinants of health. Results . Of 112 adult Asian American patients with hepatocellular cancer, 22 (20%) were Southeast Asian, and 74 (67%) described English proficiency. A score of 1 (highest risk) was observed in 1 patient (0.9%) for housing instability; 1 (0.9%) lack of transportation; no patient for financial hardship; and 1 (0.9%) for social isolation. However â€" and importantly -- total noncompletion per domain (no question answered within that domain) was observed in 90 patients (80%) for housing instability; 90 (80%) for lack of transportation; 92 (82%) for financial hardship; and 90 (80%) for social isolation. Of note, institution-wide benchmark total noncompletion rates were 0.3%, 0.3%, 47%, and 39% for these respective domains. Conclusion . High total noncompletion rates make social determinants of health data challenging to interpret and underscore the need for evidence-based guidelines on how best to capture such data in underserved patients.

Outcomes of Patients with Advanced Hepatocellular Carcinoma Receiving Lenvatinib following Immunotherapy: A Real World Evidence Study.

BACKGROUND: Lenvatinib, a multikinase inhibitor, is an FDA-approved treatment for advanced hepatocellular carcinoma (HCC) in the first-line setting. Recent trial data have established atezolizumab plus bevacizumab as well as tremelimumab plus durvalumab as preferred first-line treatment options for advanced HCC. The role of lenvatinib following progression on immunotherapy in patients with advanced HCC remains unclear. METHODS: We conducted a multicentric, retrospective analysis of patients with advanced HCC diagnosed between 2010 and 2021 at the Mayo Clinic in Minnesota, Arizona, and Florida who received immunotherapy followed by lenvatinib. Median overall survival and progression-free survival analyses were performed using the Kaplan-Meier method, and responses were determined using RECIST 1.1. Adverse events were determined using CTCAE v 4.0. RESULTS: We identified 53 patients with advanced HCC who received lenvatinib following progression on immunotherapy. Forty five (85%) patients had a Child Pugh class A at diagnosis, while 30 (58%) patients were still Child Pugh A at time of lenvatinib initiation. Lenvatinib was administered as a second-line treatment in 85% of the patients. The median PFS was 3.7 months (95% CI: 3.2-6.6), and the median OS from the time of lenvatinib initiation was 12.8 months (95% CI: 6.7-19.5). In patients with Child Pugh class A, the median OS and PFS was 14 and 5.2 months, respectively. Race, gender, and Child Pugh class was associated with OS on multivariate analysis. DISCUSSION: Our study, using real-world data, suggests that patients benefit from treatment with lenvatinib following progression on immunotherapy in advanced HCC. The optimal sequencing of therapy for patients with advanced HCC following progression on immunotherapy remains unknown, and these results need to be validated in a clinical trial.

Therapeutic Efficacy of Temsirolimus in a Patient-derived Model of Metastatic Fibrolamellar Hepatocellular Carcinoma.

BACKGROUND/AIM: Fibrolamellar hepatocellular carcinoma (FLHCC) is a rare tumor presenting in younger patients without chronic liver disease. Up to 80-100% develop recurrent disease, necessitating additional surgery or systemic treatment. Systemic options and pre-clinical treatment studies are lacking. We previously described patient-derived xenograft (PDX) development, allowing for pre-clinical studies. Herein, we develop FLHCC PDX models and utilize these to define tumor characteristics and determine the efficacy of systemic agents. MATERIALS AND METHODS: Primary and lymph node metastatic tumor tissues were obtained at the time of FLHCC resection in two patients. Tumor lysates were screened for protein upregulation. Cell lines were generated from metastatic and primary tumor tissue. The viability of the cell lines was assessed after treatment with temsirolimus, gemcitabine/oxaliplatin, and FOLFIRINOX. Two PDX models were developed from metastatic tissue. For in vivo studies, tumor-bearing mice were treated with temsirolimus, FOLFIRINOX, and Gemcitabine/oxaliplatin. RESULTS: PDX models were successfully generated from metastatic FLHCC, which closely recapitulated the original tumor. Upregulation of mTOR was seen in metastatic tissue compared to primary tumors. Cell lines from metastatic tissue demonstrated significant sensitivity to temsirolimus. In vivo testing of PDX models demonstrated a significant response to single-agent temsirolimus with minimal toxicity. CONCLUSION: Herein, we demonstrate the feasibility of developing PDX models that closely recapitulate FLHCC. Upregulation of mTOR was seen in metastatic tissue compared to primary tissue. The efficacy of mTOR inhibition with temsirolimus treatment suggests that the upregulation of the mTOR pathway may be a significant mechanism for growth in metastatic lesions and a potential target for therapeutics.

A New Look and Convergence Rate of Federated Multitask Learning With Laplacian Regularization.

Non-independent and identically distributed (non-IID) data distribution among clients is considered as the key factor that degrades the performance of federated learning (FL). Several approaches to handle non-IID data, such as personalized FL and federated multitask learning (FMTL), are of great interest to research communities. In this work, first, we formulate the FMTL problem using Laplacian regularization to explicitly leverage the relationships among the models of clients for multitask learning. Then, we introduce a new view of the FMTL problem, which, for the first time, shows that the formulated FMTL problem can be used for conventional FL and personalized FL. We also propose two algorithms and decentrali-zed () to solve the formulated FMTL problem in communication-centralized and decentralized schemes, respectively. Theoretically, we prove that the convergence rates of both algorithms achieve linear speedup for strongly convex and sublinear speedup of order 1/2 for nonconvex objectives. Experimentally, we show that our algorithms outperform the conventional algorithm FedAvg, FedProx, SCAFFOLD, and AFL in FL settings, MOCHA in FMTL settings, as well as pFedMe and Per-FedAvg in personalized FL settings.