IBD Matchmaking - Rational Combination Therapy.

A growing number of patients with Crohn's disease and ulcerative colitis have disease that is refractory to multiple advanced therapies, have undergone multiple surgeries, and require further treatment options. For this reason, there has been increasing use of multiple simultaneous advanced targeted therapies. Although the knowledge on combined advanced targeted therapy (CATT) in inflammatory bowel disease (IBD) has been largely limited to observational data and early-phase randomized controlled trials, combination of therapies is commonplace in many other diseases. This review discusses conceptual frameworks of CATT in IBD, provides context of combined therapies in other diseases, provides current evidence for CATT in IBD, and projects future applications and positioning of CATT using existing, novel, and orthogonal mechanisms of action. CATT aims to address the need to overcome low efficacy rates and frequent loss of response of current individual therapies. Both treatment exposure and disease duration are major determinants of response to therapy. Identification of safe and effective CATT may impact positioning of this strategy to apply to a broader IBD population.

Impact of Treatment Response on Risk of Serious Infections in Patients With Crohn's Disease: Secondary Analysis of the PYRAMID Registry.

BACKGROUND & AIMS: Traditional risk factors for serious infections with advanced therapies in patients with Crohn's disease (CD) have been assessed at baseline before starting therapy. We evaluated the impact of treatment response on the risk of serious infections in adalimumab-treated patients with CD through secondary analysis of the PYRAMID registry (NCT00524537). METHODS: We included patients with CD who initiated adalimumab and classified them as treatment responders (achieved steroid-free clinical remission based on patient-reported outcomes) vs nonresponders (not in steroid-free clinical remission) at 6 months after treatment initiation (landmark). We compared the risk of serious infections between responders vs nonresponders between 6 and 36 months after treatment initiation through stabilized inverse probability of treatment weighting Cox proportional hazards model. RESULTS: Of 1515 adalimumab-treated patients, 763 (50.4%) were classified as responders at 6 months (37 ± 13 y; 56% female; disease duration, 9.5 ± 8.5 y). Compared with nonresponders, responders were less likely to have moderate to severe symptoms (55.6% vs 33%), or require steroids (45.5% vs 17.3%) or opiates (6.6% vs 1.3%) at baseline, without any differences in disease location, perianal disease, and prior CD complications. During follow-up evaluation, using stabilized inverse probability of treatment weighting, responders were 34% less likely to experience serious infections compared with nonresponders (hazard ratio, 0.66; 95% CI, 0.46-0.96). Risk of gastrointestinal and extraintestinal infections was lower in responders vs nonresponders. CONCLUSIONS: Patients with CD who respond to adalimumab have a lower risk of developing serious infections compared with nonresponders. These findings underscore that initiation of advanced therapy for CD may lower the risk of serious infections through effective disease control and avoidance of corticosteroids.

Risk of Serious Adverse Gastrointestinal Events with Potassium Binders in Hospitalized Patients: A National Study.

BACKGROUND: Concerns about serious adverse gastrointestinal (GI) events with sodium polystyrene sulfonate (SPS) led to development of two new potassium binders, patiromer and sodium zirconium cyclosilicate (SZC), for treatment of hyperkalemia. OBJECTIVE: To compare risk of intestinal ischemia/thrombosis or other serious GI events associated with SPS, patiromer, or SZC in hospitalized patients. DESIGN: Retrospective cohort study. PARTICIPANTS: National sample of 3,144,960 veterans hospitalized 2016-2022 in the U.S. Department of Veterans Affairs Healthcare System. MAIN MEASURES: Demographics, comorbidities, medications and outcomes were ascertained from the VA Corporate Data Warehouse. Exposures were SPS, patiromer, SZC. Outcomes were 30-day intestinal ischemia/thrombosis, and a composite of intestinal ischemia/thrombosis, peptic ulcer/perforation or bowel resection/ostomy. KEY RESULTS: Potassium binders were used during 39,270 (1.3%) hospitalizations: SPS = 30,040 (1.0%), patiromer = 3,750 (0.1%), and SZC = 5,520 (0.2%). Intestinal ischemia/thrombosis occurred with 106/30,040 (0.4%) SPS, 12/3750 (0.3%) patiromer and 24/5520 (0.4%) SZC, vs. 6998/3,105,650 (0.2%) without potassium binder. Adjusted odds ratios (aOR) were 1.40 [95% CI, 1.16 to 1.69] with SPS, 1.36 [CI, 0.79 to 2.36] with patiromer, and 1.78 [CI, 1.21 to 2.63] with SZC exposures. Composite GI adverse events occurred with 754/30,040 (2.5%) SPS, 96/3750 (2.6%) patiromer, 2.6% SZC, vs. 144/5520 (2.4%) without binder; aOR were 1.00 [CI, 0.94 to 1.08] with SPS, 1.08 [CI, 0.89 to 1.32] with patiromer, and 1.08 [CI, 0.93 to 1.27] with SZC exposures. No statistical difference in intestinal ischemia/thrombosis between each new agent and SPS was seen (p = 0.274 for SPS vs. SZC; p = 0.916 for SPS vs. patiromer). CONCLUSION: Risk of intestinal ischemia/thrombosis or other serious adverse GI events was low and did not differ across three potassium-binding drugs.

Real-world evidence of lorlatinib therapy in Taiwanese patients with advanced anaplastic lymphoma kinase-positive non-small cell lung cancer.

BACKGROUND: Lorlatinib is a brain-penetrant, third-generation anaplastic lymphoma kinase (ALK) inhibitor indicated for ALK-positive metastatic non-small cell lung cancer (NSCLC). In a global phase II study, patients who experience disease progression despite prior treatment with ALK tyrosine kinase inhibitors (TKIs) was assessed. Herein, we report real-world clinical outcomes of lorlatinib-treated patients with ALK-positive advanced NSCLC who were heavily pretreated and progressed on first- and second-generation ALK-TKIs, in a Taiwanese population under the lorlatinib expanded access program (EAP). METHODS: This multicenter observational study examined the effectiveness and safety of ALK-positive advanced NSCLC patients that progressed from previous second-generation ALK-TKI therapy and received lorlatinib treatment subsequently. Patients who received lorlatinib treatment under EAP between Jul 2017 and Sep 2019 were eligible. Patients were followed for at least one year from the first lorlatinib treatment until study completion. RESULTS: Sixty-three patients were eligible for safety analysis (male: 46.0 %; median age: 52.8 [27.5-78.3] years; brain metastases: 81.0 %). Fifty-four patients with more than one-month lorlatinib treatment were included in the effectiveness analysis. Prior to lorlatinib treatment, 10 patients (18.5 %) received one ALK-TKI, 27 (50.0 %) received two ALK-TKIs, and 17 (31.5 %) received three or more ALK-TKIs. The overall median rwPFS was 9.2 months (95 % confidence interval: 5.3-21.1). The best overall response rate (n = 51) was 13.7 %, with a disease control rate of 80.4 %. CONCLUSION: Lorlatinib exhibits substantial activity and tolerability when used clinically in a later-line setting in a Taiwanese population with ALK-positive advanced NSCLC.

Clinical characteristics and genetic HLA marker for patients with oxaliplatin-induced adverse drug reactions.

BACKGROUND: Oxaliplatin is commonly used to treat gastrointestinal malignancies. However, its applications are limited due to potential adverse drug reactions (ADRs), particularly severe anaphylactic shock. There is no method to predict or prevent ADRs caused by oxaliplatin. Therefore, we aimed to investigate the genetic HLA predisposition and immune mechanism of oxaliplatin-induced ADRs. METHODS: A retrospective review was performed for 154 patients with ADRs induced by oxaliplatin during 2016-2021 recorded in our ADR notification system. HLA genotyping was conducted for 47 patients with oxaliplatin-induced ADRs, 1100 general population controls, and 34 oxaliplatin-tolerant controls in 2019-2023. The in vitro basophil activation test (BAT) was performed and oxaliplatin-specific IgE levels were determined. RESULTS: The incidence of oxaliplatin-induced ADRs and anaphylactic shock in our cohort was 7.1% and 0.15%, respectively. Of the 154 patients, 67.5% suffered rash/eruption; 26.0% of the patients who could not undergo oxaliplatin rechallenge were considered to show oxaliplatin-induced immune-mediated hypersensitivity reactions (HRs). The genetic study found that the HLA-DRB∗12:01 allele was associated with oxaliplatin-induced HRs compared to the general population controls (sensitivity = 42.9%; odds ratio [OR] = 3.4; 95% CI = 1.4-8.2; P = 0.008) and tolerant controls (OR = 12; 95% CI = 2.3-63.7; P = 0.001). The in vitro BAT showed higher activation of CD63+ basophils in patients with oxaliplatin-induced HRs compared to the tolerant controls (P < 0.05). Only four patients (8.5%) with oxaliplatin-induced ADRs were positive for oxaliplatin-specific IgE. CONCLUSIONS: This study found that 26.0% of patients with oxaliplatin-induced ADRs could not undergo oxaliplatin rechallenge. HLA-DRB∗12:01 is regarded as a genetic marker for oxaliplatin-induced hypersensitivity.

Comparative Analysis of Tolerance to Experimentally Induced Astigmatism with Three Types of Multifocal Intraocular Lenses.

Purpose: The effect of residual astigmatism and its axis on distance and near visual acuities (VAs) with multifocal intraocular lenses (IOLs) has not been studied extensively. This study compared the tolerance to experimentally induced residual astigmatism among bifocal, trifocal, and extended depth-of-focus (EDOF) IOLs. Patients and Methods: This retrospective, comparative study included 70 eyes of 70 patients implanted with bifocal, trifocal, or EDOF IOLs. Distance and near VAs were assessed with experimentally induced astigmatism by placing positive cylindrical lenses in increments of 0.50 diopters to 2.00 diopters at 90° and 180° axes over the best distance correction. Results: Both distance and near VAs worsened with increasing magnitudes of experimentally induced astigmatism except in the EDOF group, in which the near VA remained within a clinically acceptable limit, ie, within one line from the best corrected VA under all ranges of experimentally induced astigmatism. Furthermore, the EDOF group showed the highest astigmatic threshold for losing VA lines following experimental astigmatic induction at both distance and near. The distance VA was generally better at with-the-rule (WTR) than against-the-rule (ATR) astigmatism for all three IOL groups. On the other hand, the near VA was generally better at WTR than ATR astigmatism in the bifocal group, comparable between WTR and ATR astigmatism in the trifocal group, and generally better at ATR than WTR astigmatism in the EDOF group. Conclusion: The EDOF IOL demonstrated the highest tolerance to experimentally induced astigmatism at both distance and near. VA was generally less affected by WTR astigmatism than ATR astigmatism, especially at distance. We proposed the residual astigmatism thresholds for clinically acceptable VA reduction in all three IOL groups.

Distinct changes to pancreatic volume rather than pancreatic autoantibody positivity: insights into immune checkpoint inhibitors induced diabetes mellitus.

BACKGROUND: Immune checkpoint inhibitors (ICI) are promising treatment options for various cancers. However, their use is associated with immune-related adverse events (irAEs), including ICI-induced diabetes mellitus (ICI-DM). This study aimed to investigate the clinical features of ICI-DM, with a particular focus on alterations to pancreatic volume. METHODS: We conducted a retrospective review of 2829 patients who received ICI treatment at the Chang Gung Memorial Hospital, Linkou, between January 2014 and December 2021. New-onset diabetes or diabetic ketoacidosis (DKA) was identified in ten patients receiving ICI therapy. Pancreatic volumes were assessed by manual segmentation of computed tomography (CT) images before and after ICI-DM diagnosis. RESULTS: Among these ten patients, nivolumab was the most commonly used ICI (50.0%), followed by pembrolizumab (30.0%) and atezolizumab (20.0%). One patient received combination therapy with nivolumab and ipilimumab. The median age was 63.01 years (range: 40.1 - 87.8). ICI-DM developed after a median of 13.5 cycles (range: 2 - 42) of ICI treatment or 9.85 months (range:1.5 - 21.3) since ICI initiation. The initial presentation was DKA in 60.0% of patients. All patients had low or undetectable C-peptide levels (range: <0.033 - 0.133 nmol/L) and were negative for most type 1 diabetes mellitus (T1DM)-related autoantibodies; only one patient tested positive for glutamic acid decarboxylase antibodies. CT imaging revealed significant pancreatic atrophy, with a median pancreatic volume decrease of 19.92% (P = 0.038) from baseline and sustained significant decline at last follow-up (median - 37.14%, P = 0.012). CONCLUSIONS: ICI-DM is often accompanied by pancreatic atrophy and approximately two-thirds of patients initially present with DKA. Although the majority of ICI-DM patients lack T1DM-related autoantibodies, identifying diminished pancreatic volumes through CT imaging provides valuable clues into the subclinical aspects of ICI-DM development, aiding in the prevention of diabetic emergencies. TRIAL REGISTRATION: Not applicable.

Automated Spatial Omics Landscape Analysis Approach Reveals Novel Tissue Architectures in Ulcerative Colitis.

The utility of spatial omics in leveraging cellular interactions in normal and diseased states for precision medicine is hampered by a lack of strategies for matching disease states with spatial heterogeneity-guided cellular annotations. Here we use a spatial context-dependent approach that matches spatial pattern detection to cell annotation. Using this approach in existing datasets from ulcerative colitis patient colonic biopsies, we identified architectural complexities and associated difficult-to-detect rare cell types in ulcerative colitis germinal-center B cell follicles. Our approach deepens our understanding of health and disease pathogenesis, illustrates a strategy for automating nested architecture detection for highly multiplexed spatial biology data, and informs precision diagnosis and therapeutic strategies.

Effectiveness and safety of afatinib, gefitinib, and erlotinib for treatment-naïve elderly patients with epidermal growth factor receptor-mutated advanced non-small-cell lung cancer: a multi-institute retrospective study.

BACKGROUND: In real-world practice, most patients with lung cancer are diagnosed when they are aged ≥65 years. However, clinical trials tend to lack data for the elderly population. Therefore, we aimed to describe the effectiveness and safety of afatinib, gefitinib, and erlotinib for elderly patients with epidermal growth factor receptor (EGFR)-mutated advanced non-small-cell lung cancer (NSCLC). METHODS: Treatment-naïve patients with EGFR-mutated advanced NSCLC were enrolled at many hospitals in Taiwan. Patient characteristics and the effectiveness and safety of afatinib, gefitinib, and erlotinib were compared. RESULTS: This study enrolled 1,343 treatment-naïve patients with EGFR-mutated advanced NSCLC, of whom 554 were aged <65 years, 383 were aged 65-74 years, 323 were aged 75-84 years, and 83 were aged ≥85 years. For elderly patients, afatinib was more effective, with a median progression-free survival (PFS) of 14.7 months and overall survival (OS) of 22.2 months, than gefitinib (9.9 months and 17.7 months, respectively) and erlotinib (10.8 months and 18.5 months, respectively; PFS: p = 0.003; OS: p = 0.026). However, grade ≥3 adverse events, including skin toxicities, paronychia, mucositis, and diarrhea, were more frequently experienced by patients receiving afatinib than those receiving gefitinib or erlotinib. CONCLUSIONS: This large retrospective study provides real-world evidence of the effectiveness and safety of EGFR-TKIs for elderly patients with EGFR-mutated advanced NSCLC, a population that is often underrepresented in clinical trials and real-world evidence. Afatinib was more effective as a first-line treatment than gefitinib or erlotinib for elderly patients with EGFR-mutated advanced NSCLC.

Automated spatial omics landscape analysis approach reveals novel tissue architectures in ulcerative colitis.

The utility of spatial omics in leveraging cellular interactions in normal and diseased states for precision medicine is hampered by a lack of strategies for matching disease states with spatial heterogeneity-guided cellular annotations. Here we use a spatial context-dependent approach that matches spatial pattern detection to cell annotation. Using this approach in existing datasets from ulcerative colitis patient colonic biopsies, we identified architectural complexities and associated difficult-to-detect rare cell types in ulcerative colitis germinal-center B cell follicles. Our approach deepens our understanding of health and disease pathogenesis, illustrates a strategy for automating nested architecture detection for highly multiplexed spatial biology data, and informs precision diagnosis and therapeutic strategies.

Proton Pencil Beam Scanning Facilitates the Safe Treatment of Extended Radiation Targets for Hodgkin Lymphoma: A Report from the Proton Collaborative Group Registry.

Because proton beam therapy (PBT) can lower the dose of radiation to the heart, lungs, and breast, it is an established radiation modality for patients with Hodgkin lymphoma (HL). Pencil beam scanning (PBS) PBT facilitates the treatment of more extensive targets. This may be especially of value for lymphoma patients who require RT to both mediastinal and axillary targets, defined here as extended target RT (ETRT), given the target distribution and need to minimize the lung, heart, and breast dose. Using the Proton Collaborative Group registry, we identified patients with HL treated with PBT to both their mediastinum and axilla, for which DICOM-RT was available. All patients were treated with PBS. To evaluate the dosimetric impact of PBS, we compared delivered PBS plans with VMAT butterfly photon plans optimized to have the same target volume coverage, when feasible. Between 2016 and 2021, twelve patients (median 26 years) received PBS ETRT (median 30.6 Gy (RBE)). Despite the large superior/inferior (SI, median 22.2 cm) and left/right (LR, median 22.8 cm) extent of the ETRT targets, all patients were treated with one isocenter except for two patients (both with SI and LR > 30 cm). Most commonly, anterior beams, with or without posterior beams, were used. Compared to photons, PBS had greater target coverage, better conformity, and lower dose heterogeneity while achieving lower doses to the lungs and heart, but not to the breast. No acute grade 3+ toxicities were reported, including pneumonitis. Proton ETRT in this small cohort was safely delivered with PBS and was associated with an improved sparing of the heart and lungs compared to VMAT.

Reirradiation With Proton Therapy for Recurrent Malignancies of the Esophagus and Gastroesophageal Junction: Results of the Proton Collaborative Group Multi-Institutional Prospective Registry Trial.

Purpose: Treatment options for recurrent esophageal cancer (EC) previously treated with radiation therapy (RT) are limited. Reirradiation (reRT) with proton beam therapy (PBT) can offer lower toxicities by limiting doses to surrounding tissues. In this study, we present the first multi-institutional series reporting on toxicities and outcomes after reRT for locoregionally recurrent EC with PBT. Methods and Materials: Analysis of the prospective, multicenter, Proton Collaborative Group registry of patients with recurrent EC who had previously received photon-based RT and underwent PBT reRT was performed. Patient/tumor characteristics, treatment details, outcomes, and toxicities were collected. Local control (LC), distant metastasis-free survival (DMFS), and overall survival (OS) were estimated using the Kaplan-Meier method. Event time was determined from reRT start. Results: Between 2012 and 2020, 31 patients received reRT via uniform scanning/passive scattering (61.3%) or pencil beam scanning (38.7%) PBT at 7 institutions. Median prior RT, PBT reRT, and cumulative doses were 50.4 Gy (range, 37.5-110.4), 48.6 Gy (relative biological effectiveness) (25.2-72.1), and 99.9 Gy (79.1-182.5), respectively. Of these patients, 12.9% had 2 prior RT courses, and 67.7% received PBT with concurrent chemotherapy. Median follow-up was 7.2 months (0.9-64.7). Post-PBT, there were 16.7% locoregional only, 11.1% distant only, and 16.7% locoregional and distant recurrences. Six-month LC, DMFS, and OS were 80.5%, 83.4%, and 69.1%, respectively. One-year LC, DMFS, and OS were 67.1%, 83.4%, and 27%, respectively. Acute grade ≥3 toxicities occurred in 23% of patients, with 1 acute grade 5 toxicity secondary to esophageal hemorrhage, unclear if related to reRT or disease progression. No grade ≥3 late toxicities were reported. Conclusions: In the largest report to date of PBT for reRT in patients with recurrent EC, we observed acceptable acute toxicities and encouraging rates of disease control. However, these findings are limited by the poor prognoses of these patients, who are at high risk of mortality. Further research is needed to better assess the long-term benefits and toxicities of PBT in this specific patient population.

Stereotactic Body Proton Therapy Versus Conventionally Fractionated Proton Therapy for Early Prostate Cancer: A Randomized, Controlled, Phase 3 Trial.

PURPOSE: We aimed to determine if ultra-hypofractionated proton therapy delivered via stereotactic body proton therapy (SBPT) is non-inferior to conventionally fractionated proton therapy (CFPT) in patients with early prostate cancer. MATERIALS AND METHODS: This study was a multicenter, randomized, controlled, non-inferiority phase 3 trial that included patients with histologically confirmed low-risk prostate adenocarcinoma defined by Gleason score grouping 1, PSA <10 ng/mL, and clinical stage T1-2a N0 M0 according to AJCC 7th ed. Eligible participants were randomly assigned initially at a 1:1 ratio and later at a 2:1 ratio to SBPT (38 Gy in 5 fractions) or CFPT (79.2 Gy in 44 fractions). The primary endpoint was freedom from failure (FFF) at 2 years from the date of randomization. Non-inferiority for FFF was determined based on one-sided confidence intervals. Toxicities were compared at different time points using Fisher's Exact test. Health-related quality-of-life (HRQoL) was analyzed at different time points using a mixed-effects linear model. This trial is registered with ClinicalTrials.gov, NCT01230866, and is closed to accrual. RESULTS: Between December 10, 2010, and September 29, 2020, 144 patients were enrolled and 135 were randomly assigned (90 to the SBPT group and 45 to the CFPT group). The median follow-up was 5 years (IQR 3.9-5.2). The 2-year FFF was 100% for both groups, with the one-sided 5-year risk difference in FFF between groups reported as 2.63% (90% CI: -1.70%-6.96%), favoring the SBRT arm, thus fulfilling the pre-specified criteria for non-inferiority of SBPT compared to CFPT. Rates of gastrointestinal (GI) and genitourinary (GU) G2 and G3 toxicities did not differ significantly between groups but the the study was not powered to detect significant toxicity differences. Also, HRQoL metrics did not differ significantly between groups over the study median follow up. CONCLUSIONS: SBPT is non-inferior to CFPT regarding FFF, with similar long-term GU and GI toxicity rates and minimal impact in patient reported HRQoL over time.

Functional Diversity of Memory CD8 T Cells is Spatiotemporally Imprinted.

Tissue-resident memory CD8 T cells (TRM) kill infected cells and recruit additional immune cells to limit pathogen invasion at barrier sites. Small intestinal (SI) TRM cells consist of distinct subpopulations with higher expression of effector molecules or greater memory potential. We hypothesized that occupancy of diverse anatomical niches imprints these distinct TRM transcriptional programs. We leveraged human samples and a murine model of acute systemic viral infection to profile the location and transcriptome of pathogen-specific TRM cell differentiation at single-transcript resolution. We developed computational approaches to capture cellular locations along three anatomical axes of the small intestine and to visualize the spatiotemporal distribution of cell types and gene expression. TRM populations were spatially segregated: with more effector- and memory-like TRM preferentially localized at the villus tip or crypt, respectively. Modeling ligand-receptor activity revealed patterns of key cellular interactions and cytokine signaling pathways that initiate and maintain TRM differentiation and functional diversity, including different TGFß sources. Alterations in the cellular networks induced by loss of TGFßRII expression revealed a model consistent with TGFß promoting progressive TRM maturation towards the villus tip. Ultimately, we have developed a framework for the study of immune cell interactions with the spectrum of tissue cell types, revealing that T cell location and functional state are fundamentally intertwined.

Gut microbiota and clinical response to immune checkpoint inhibitor therapy in patients with advanced cancer.

BACKGROUND: There is currently no well-accepted consensus on the association between gut microbiota and the response to treatment of immune checkpoint inhibitors (ICIs) in patients with advanced cancer. METHODS: Fecal samples were collected before ICI treatment. Gut microbiota was analyzed using 16 S ribosomal RNA sequencing. We investigated the relationship between the α-diversity of fecal microbiota and patients' clinical outcomes. Microbiota profiles from patients and healthy controls were determined. Pre-treatment serum was examined by cytokine array. RESULTS: We analyzed 74 patients, including 42 with melanoma, 8 with kidney cancer, 13 with lung cancer, and 11 with other cancers. Combination therapy of anti-PD1 and anti-CTLA-4 was used in 14 patients, and monotherapy in the rest. Clinical benefit was observed in 35 (47.3 %) cases, including 2 complete responses, 16 partial responses, and 17 stable diseases according to RECIST criteria. No significant difference in α-diversity was found between the benefiter and non-benefiter groups. However, patients with α-diversity within the range of our healthy control had a significantly longer median overall survival (18.9 months), compared to the abnormal group (8.2 months) (p = 0.041, hazard ratio = 0.546) for all patients. The microbiota composition of the benefiters was similar to that of healthy individuals. Furthermore, specific bacteria, such as Prevotella copri and Faecalibacterium prausnitzii, were associated with a favorable outcome. We also observed that serum IL-18 before treatment was significantly lower in the benefiters, compared to non-benefiters. CONCLUSIONS: The α-diversity of gut microbiota is positively correlated with more prolonged overall survival in cancer patients following ICI therapy.

Evolution of Proton Radiotherapy Brainstem Constraints on the Pediatric Proton/Photon Consortium Registry.

INTRODUCTION: Increasing concern that brainstem toxicity incidence after proton radiotherapy (PRT) might be higher than with photons led to a 2014 XXXX (XX) landmark paper identifying its risk factors and proposing more conservative dose constraints. We evaluated how practice patterns changed among the XXXX (XXXX). METHODS: This prospective multicenter cohort study gathered data from patients under age 22 enrolled on the XXXX, treated between 2002-2019 for primary posterior fossa brain tumors. After standardizing brainstem contours, we garnered dosimetry data and correlated those meeting the 2014 proton-specific brainstem constraint guidelines by treatment era, histology, and extent of surgical resection. RESULTS: A total of 467 patients with evaluable PRT plans were reviewed. Median age was 7.1 years (range: <1-21.9), 63.0% (n=296) were male, 76.0% (n=357) were white, and predominant histologies were medulloblastoma (55.0%, n=256) followed by ependymoma (27.0%, n=125). Extent of resection was mainly gross total resection (GTR) (67.0%, n=312), followed by subtotal resection (STR) or biopsy (20.0%, n=92). The XX brainstem constraint metrics most often exceeded were the goal D50% of 52.4 GyRBE (43.3%, n=202) and maximal D50% of 54 GyRBE (12.6%, n=59). The compliance rate increased after the new guidelines (2002-2014: 64.0% vs. 2015-2019: 74.6%, p=0.02), except for ependymoma (46.3% pre vs. 50.0% post guidelines, p=0.86), presenting lower compliance (48.8%) in comparison to medulloblastoma/PNET/pineoblastoma (77.7%), glioma (89.1%), and ATRT (90.9%) (p<0.001). Degree of surgical resection did not affect compliance rates (GTR/NTR 71.0% vs. STR/biopsy 72.8%, p=0.45), even within the ependymoma subset (GTR/NTR 50.5% vs. STR/biopsy 38.1%, p=0.82). CONCLUSION: Since the publication of the XX guidelines, the pediatric proton community has implemented more conservative brainstem constraints in all patients except those with ependymoma, irrespective of residual disease after surgery. Future work will evaluate if this change in practice is associated with decreased rates of brainstem toxicity.

Methods for improving colorectal cancer annotation efficiency for artificial intelligence-observer training.

BACKGROUND: Missing occult cancer lesions accounts for the most diagnostic errors in retrospective radiology reviews as early cancer can be small or subtle, making the lesions difficult to detect. Second-observer is the most effective technique for reducing these events and can be economically implemented with the advent of artificial intelligence (AI). AIM: To achieve appropriate AI model training, a large annotated dataset is necessary to train the AI models. Our goal in this research is to compare two methods for decreasing the annotation time to establish ground truth: Skip-slice annotation and AI-initiated annotation. METHODS: We developed a 2D U-Net as an AI second observer for detecting colorectal cancer (CRC) and an ensemble of 5 differently initiated 2D U-Net for ensemble technique. Each model was trained with 51 cases of annotated CRC computed tomography of the abdomen and pelvis, tested with 7 cases, and validated with 20 cases from The Cancer Imaging Archive cases. The sensitivity, false positives per case, and estimated Dice coefficient were obtained for each method of training. We compared the two methods of annotations and the time reduction associated with the technique. The time differences were tested using Friedman's two-way analysis of variance. RESULTS: Sparse annotation significantly reduces the time for annotation particularly skipping 2 slices at a time (P < 0.001). Reduction of up to 2/3 of the annotation does not reduce AI model sensitivity or false positives per case. Although initializing human annotation with AI reduces the annotation time, the reduction is minimal, even when using an ensemble AI to decrease false positives. CONCLUSION: Our data support the sparse annotation technique as an efficient technique for reducing the time needed to establish the ground truth.