Explore the vast range of titles available.

Artificial intelligence (AI) can predict the presence of molecular alterations directly from routine histopathology slides. However, training robust AI systems requires large datasets for which data collection faces practical, ethical and legal obstacles. These obstacles could be overcome with swarm learning (SL), in which partners jointly train AI models while avoiding data transfer and monopolistic data governance. Here, we demonstrate the successful use of SL in large, multicentric datasets of gigapixel histopathology images from over 5,000 patients. We show that AI models trained using SL can predict BRAF mutational status and microsatellite instability directly from hematoxylin and eosin (H&E)-stained pathology slides of colorectal cancer. We trained AI models on three patient cohorts from Northern Ireland, Germany and the United States, and validated the prediction performance in two independent datasets from the United Kingdom. Our data show that SL-trained AI models outperform most locally trained models, and perform on par with models that are trained on the merged datasets. In addition, we show that SL-based AI models are data efficient. In the future, SL can be used to train distributed AI models for any histopathology image analysis task, eliminating the need for data transfer. A decentralized, privacy-preserving machine learning framework used to train a clinically relevant AI system identifies actionable molecular alterations in patients with colorectal cancer by use of routine histopathology slides collected in real-world settings.

Gastric cancer is the fifth most common cancer and the third most common cause of cancer death globally. Risk factors for the condition include Helicobacter pylori infection, age, high salt intake, and diets low in fruit and vegetables. Gastric cancer is diagnosed histologically after endoscopic biopsy and staged using CT, endoscopic ultrasound, PET, and laparoscopy. It is a molecularly and phenotypically highly heterogeneous disease. The main treatment for early gastric cancer is endoscopic resection. Non-early operable gastric cancer is treated with surgery, which should include D2 lymphadenectomy (including lymph node stations in the perigastric mesentery and along the celiac arterial branches). Perioperative or adjuvant chemotherapy improves survival in patients with stage 1B or higher cancers. Advanced gastric cancer is treated with sequential lines of chemotherapy, starting with a platinum and fluoropyrimidine doublet in the first line; median survival is less than 1 year. Targeted therapies licensed to treat gastric cancer include trastuzumab (HER2-positive patients first line), ramucirumab (anti-angiogenic second line), and nivolumab or pembrolizumab (anti-PD-1 third line).

Microsatellite instability determines whether patients with gastrointestinal cancer respond exceptionally well to immunotherapy. However, in clinical practice, not every patient is tested for MSI, because this requires additional genetic or immunohistochemical tests. Here we show that deep residual learning can predict MSI directly from H&E histology, which is ubiquitously available. This approach has the potential to provide immunotherapy to a much broader subset of patients with gastrointestinal cancer.A deep residual learning framework identifies microsatellite instability in histology slides from patients with cancer and can be used to guide immunotherapy.

LGR5 marks resident adult epithelial stem cells at the gland base in the mouse pyloric stomach 1 , but the identity of the equivalent human stem cell population remains unknown owing to a lack of surface markers that facilitate its prospective isolation and validation. In mouse models of intestinal cancer, LGR5 + intestinal stem cells are major sources of cancer following hyperactivation of the WNT pathway 2 . However, the contribution of pyloric LGR5 + stem cells to gastric cancer following dysregulation of the WNT pathway—a frequent event in gastric cancer in humans 3 —is unknown. Here we use comparative profiling of LGR5 + stem cell populations along the mouse gastrointestinal tract to identify, and then functionally validate, the membrane protein AQP5 as a marker that enriches for mouse and human adult pyloric stem cells. We show that stem cells within the AQP5 + compartment are a source of WNT-driven, invasive gastric cancer in vivo, using newly generated Aqp5-creERT2 mouse models. Additionally, tumour-resident AQP5 + cells can selectively initiate organoid growth in vitro, which indicates that this population contains potential cancer stem cells. In humans, AQP5 is frequently expressed in primary intestinal and diffuse subtypes of gastric cancer (and in metastases of these subtypes), and often displays altered cellular localization compared with healthy tissue. These newly identified markers and mouse models will be an invaluable resource for deciphering the early formation of gastric cancer, and for isolating and characterizing human-stomach stem cells as a prerequisite for harnessing the regenerative-medicine potential of these cells in the clinic. AQP5 is identified as a marker for pyloric stem cells in humans and mice, and stem cells in the AQP5 + compartment are shown to be a source of invasive gastric cancer in mouse models.

Background 10–20% of patients with gastric cancer (GC) have HER2+ tumors. Addition of trastuzumab (T) to cisplatin/fluoropyrimidine-based chemotherapy (CT) improved survival in metastatic, HER2+ GC. When pertuzumab (P) was added to neoadjuvant T and CT, a significant increase in histopathological complete response rate was observed in HER2+ breast cancer. This study aims to investigate the added benefit of using both HER2 targeting drugs (T alone or the combination of T + P), in combination with perioperative CT for localized HER2+ GC. Methods This is a prospective, randomized, open-label, phase II trial. HER2 status from patients with resectable GC (UICC TNM7 tumor stage Ib-III) will be centrally determined. Two hundred and-fifteen patients from 52 sites in 14 countries will be centrally randomized (1:2:2 ratio) to one of the following treatment arms: Standard: CT alone. CT regimens will be FLOT (5-FU, leucovorin, oxaliplatin, taxotere) CapOx (capecitabine, oxaliplatin) or FOLFOX (5-FU, leucovorin, oxaliplatin) according to investigator’s choice in Europe, and cisplatin/capecitabine in Asia. Experimental arm 1 : CT as in control group, plus T (8 mg/kg loading dose, followed by 6 mg/kg every 3 weeks) at day 1, independent of CT chosen for 3 cycles of 3 weeks before and after surgery. Experimental arm 2: CT plus T as in experimental arm 1, plus P (840 mg every 3 weeks) on day 1. Adjuvant treatment with T or T + P will continue for 17 cycles in total. Stratification factors are: histology (intestinal/non-intestinal); region (Asia vs Europe); location (GEJ vs non-GEJ); HER2 immunohistochemistry score (IHC 3+ vs IHC 2+/FISH+) and chemotherapy regimen. Primary objective is to detect an increase in the major pathological response rate from 25 to 45% either with CT plus T alone, or with CT plus the combination of T and P. Discussion Depending on the results of the INNOVATION trial, the addition of HER2 targeted treatment with either T or T and P to CT may inform future study designs or become a standard in the perioperative management HER2+ GC. Trial registration This article reports a health care intervention on human participants and was registered on July 10, 2014 under ClinicalTrials.gov identifier: NCT02205047 ; EudraCT: 2014–000722-38.

Neoadjuvant chemotherapy before surgery improves survival compared with surgery alone for patients with oesophageal cancer. The OE05 trial assessed whether increasing the duration and intensity of neoadjuvant chemotherapy further improved survival compared with the current standard regimen. OE05 was an open-label, phase 3, randomised clinical trial. Patients with surgically resectable oesophageal adenocarcinoma classified as stage cT1N1, cT2N1, cT3N0/N1, or cT4N0/N1 were recruited from 72 UK hospitals. Eligibility criteria included WHO performance status 0 or 1, adequate respiratory, cardiac, and liver function, white blood cell count at least 3 × 109 cells per L, platelet count at least 100 × 109 platelets per L, and a glomerular filtration rate at least 60 mL/min. Participants were randomly allocated (1:1) using a computerised minimisation program with a random element and stratified by centre and tumour stage, to receive two cycles of cisplatin and fluorouracil (CF; two 3-weekly cycles of cisplatin [80 mg/m2 intravenously on day 1] and fluorouracil [1 g/m2 per day intravenously on days 1–4]) or four cycles of epirubicin, cisplatin, and capecitabine (ECX; four 3-weekly cycles of epirubicin [50 mg/m2] and cisplatin [60 mg/m2] intravenously on day 1, and capecitabine [1250 mg/m2] daily throughout the four cycles) before surgery, stratified according to centre and clinical disease stage. Neither patients nor study staff were masked to treatment allocation. Two-phase oesophagectomy with two-field (abdomen and thorax) lymphadenectomy was done within 4–6 weeks of completion of chemotherapy. The primary outcome measure was overall survival, and primary and safety analyses were done in the intention-to-treat population. This trial is registered with the ISRCTN registry (number 01852072) and ClinicalTrials.gov (NCT00041262), and is completed. Between Jan 13, 2005, and Oct 31, 2011, 897 patients were recruited and 451 were assigned to the CF group and 446 to the ECX group. By Nov 14, 2016, 327 (73%) of 451 patients in the CF group and 302 (68%) of 446 in the ECX group had died. Median survival was 23·4 months (95% CI 20·6–26·3) with CF and 26·1 months (22·5–29·7) with ECX (hazard ratio 0·90 (95% CI 0·77–1·05, p=0·19). No unexpected chemotherapy toxicity was seen, and neutropenia was the most commonly reported event (grade 3 or 4 neutropenia: 74 [17%] of 446 patients in the CF group vs 101 [23%] of 441 people in the ECX group). The proportions of patients with postoperative complications (224 [56%] of 398 people for whom data were available in the CF group and 233 [62%] of 374 in the ECX group; p=0·089) were similar between the two groups. One patient in the ECX group died of suspected treatment-related neutropenic sepsis. Four cycles of neoadjuvant ECX compared with two cycles of CF did not increase survival, and cannot be considered standard of care. Our study involved a large number of centres and detailed protocol with comprehensive prospective assessment of health-related quality of life in a patient population confined to people with adenocarcinomas of the oesophagus and gastro-oesophageal junction (Siewert types 1 and 2). Alternative chemotherapy regimens and neoadjuvant chemoradiation are being investigated to improve outcomes for patients with oesophageal carcinoma. Cancer Research UK and Medical Research Council Clinical Trials Unit at University College London.

Peri-operative chemotherapy and surgery is a standard of care for patients with resectable oesophagogastric adenocarcinoma. Bevacizumab, a monoclonal antibody against VEGF, improves the proportion of patients responding to treatment in advanced gastric cancer. We aimed to assess the safety and efficacy of adding bevacizumab to peri-operative chemotherapy in patients with resectable gastric, oesophagogastric junction, or lower oesophageal adenocarcinoma. In this multicentre, randomised, open-label phase 2–3 trial, we recruited patients aged 18 years and older with histologically proven, resectable oesophagogastric adenocarcinoma from 87 UK hospitals and cancer centres. We randomly assigned patients 1:1 to receive peri-operative epirubicin, cisplatin, and capecitabine chemotherapy or chemotherapy plus bevacizumab, in addition to surgery. Patients in the control group (chemotherapy alone) received three pre-operative and three post-operative cycles of epirubicin, cisplatin, and capecitabine chemotherapy: 50 mg/m2 epirubicin and 60 mg/m2 cisplatin on day 1 and 1250 mg/m2 oral capecitabine on days 1–21. Patients in the investigational group received the same treatment as the control group plus 7·5 mg/kg intravenous bevacizumab on day 1 of every cycle of chemotherapy and for six further doses once every 21 days following chemotherapy, as maintenance treatment. Randomisation was done by means of a telephone call to the Medical Research Council Clinical Trials Unit, where staff used a computer programme that implemented a minimisation algorithm with a random element to establish the allocation for the patient at the point of randomisation. Patients were stratified by chemotherapy centre, site of tumour, and tumour stage. The primary outcome for the phase 3 stage of the trial was overall survival (defined as the time from randomisation until death from any cause), analysed in the intention-to-treat population. Here, we report the primary analysis results of the trial; all patients have completed treatment and the required number of primary outcome events has been reached. This study is registered as an International Standard Randomised Controlled Trial, number ISRCTN 46020948, and with ClinicalTrials.gov, number NCT00450203. Between Oct 31, 2007, and March 25, 2014, 1063 patients were enrolled and randomly assigned to receive chemotherapy alone (n=533) or chemotherapy plus bevacizumab (n=530). At the time of analysis, 508 deaths were recorded (248 in the chemotherapy alone group and 260 in the chemotherapy plus bevacizumab group). 3-year overall survival was 50·3% (95% CI 45·5–54·9) in the chemotherapy alone group and 48·1% (43·2–52·7) in the chemotherapy plus bevacizumab group (hazard ratio [HR] 1·08, 95% CI 0·91–1·29; p=0·36). Apart from neutropenia no other toxic effects were reported at grade 3 or worse severity in more than 10% of patients in either group. Wound healing complications were more prevalent in the bevacizumab group, occurring in 53 (12%) patients in this group compared with 33 (7%) patients in the chemotherapy alone group. In patients who underwent oesophagogastrectomy, post-operative anastomotic leak rates were higher in the chemotherapy plus bevacizumab group (23 [10%] of 233 in the chemotherapy alone group vs 52 [24%] of 220 in the chemotherapy plus bevacizumab group); therefore, recruitment of patients with lower oesophageal or junctional tumours planned for an oesophagogastric resection was stopped towards the end of the trial. Serious adverse events for all patients included anastomotic leaks (30 events in chemotherapy alone group vs 69 in the chemotherapy plus bevacizumab group), and infections with normal neutrophil count (42 events vs 53). The results of this trial do not provide any evidence for the use of bevacizumab in combination with peri-operative epiribicin, cisplatin, and capecitabine chemotherapy for patients with resectable gastric, oesophagogastric junction, or lower oesophageal adenocarcinoma. Bevacizumab might also be associated with impaired wound healing. Cancer Research UK, MRC Clinical Trials Unit at University College London, and F Hoffmann-La Roche Limited.

Background Combination of immunotherapy and chemotherapy is recommended for first line treatment of gastric adenocarcinoma (GC) patients with locally advanced unresectable disease or metastatic disease. However, data regarding the concordance rate between PD-L1 combined positive score (CPS) in primary GC and matched regional lymph node metastasis (LNmet) or matched distant metastasis (Dmet) is limited. Methods Tissue microarray sections from primary resected GC, LNmet and Dmet were immunohistochemically stained with anti-PD-L1 (clone SP263). PD-L1 expression was scored separately in tumour cells and immune cells and compared between matched primary GC, LNmet and/or Dmet. CPS was calculated and results for CPS cut-offs 1 and 5 were compared between matched samples. Results 275 PD-L1 stained GC were analysed. 189 primary GC had matched LNmet. CPS cut-off 1 concordance rate between primary GC and LNmet was 77%. 23 primary GC had matched Dmet but no matched LNmet, CPS cut-off 1 concordance rate was 70%. 63 primary GC had both matched LNmet and matched Dmet, CPS cut-off 1 concordance rate of 67%. CPS cut-off 5 results were similar. The proportion of PD-L1 positive tumour cells increased from primary GC (26%) to LNmet (42%) and was highest in Dmet (75%). Conclusion Our study showed up to 33% discordance of PD-L1 CPS between primary GC and LNmet and/or Dmet suggesting that multiple biopsies of primary GC and metastatic sites might need to be tested before considering treatment options. Moreover, this is the first study that seems to suggest that tumour cells acquire PD-L1 expression during disease progression.

Background Computational pathology uses deep learning (DL) to extract biomarkers from routine pathology slides. Large multicentric datasets improve performance, but such datasets are scarce for gastric cancer. This limitation could be overcome by Swarm Learning (SL). Methods Here, we report the results of a multicentric retrospective study of SL for prediction of molecular biomarkers in gastric cancer. We collected tissue samples with known microsatellite instability (MSI) and Epstein–Barr Virus (EBV) status from four patient cohorts from Switzerland, Germany, the UK and the USA, storing each dataset on a physically separate computer. Results On an external validation cohort, the SL-based classifier reached an area under the receiver operating curve (AUROC) of 0.8092 (± 0.0132) for MSI prediction and 0.8372 (± 0.0179) for EBV prediction. The centralized model, which was trained on all datasets on a single computer, reached a similar performance. Conclusions Our findings demonstrate the feasibility of SL-based molecular biomarkers in gastric cancer. In the future, SL could be used for collaborative training and, thus, improve the performance of these biomarkers. This may ultimately result in clinical-grade performance and generalizability.

ObjectiveTo date, there are no predictive biomarkers to guide selection of patients with gastric cancer (GC) who benefit from paclitaxel. Stomach cancer Adjuvant Multi-Institutional group Trial (SAMIT) was a 2×2 factorial randomised phase III study in which patients with GC were randomised to Pac-S-1 (paclitaxel +S-1), Pac-UFT (paclitaxel +UFT), S-1 alone or UFT alone after curative surgery.DesignThe primary objective of this study was to identify a gene signature that predicts survival benefit from paclitaxel chemotherapy in GC patients. SAMIT GC samples were profiled using a customised 476 gene NanoString panel. A random forest machine-learning model was applied on the NanoString profiles to develop a gene signature. An independent cohort of metastatic patients with GC treated with paclitaxel and ramucirumab (Pac-Ram) served as an external validation cohort.ResultsFrom the SAMIT trial 499 samples were analysed in this study. From the Pac-S-1 training cohort, the random forest model generated a 19-gene signature assigning patients to two groups: Pac-Sensitive and Pac-Resistant. In the Pac-UFT validation cohort, Pac-Sensitive patients exhibited a significant improvement in disease free survival (DFS): 3-year DFS 66% vs 40% (HR 0.44, p=0.0029). There was no survival difference between Pac-Sensitive and Pac-Resistant in the UFT or S-1 alone arms, test of interaction p<0.001. In the external Pac-Ram validation cohort, the signature predicted benefit for Pac-Sensitive (median PFS 147 days vs 112 days, HR 0.48, p=0.022).ConclusionUsing machine-learning techniques on one of the largest GC trials (SAMIT), we identify a gene signature representing the first predictive biomarker for paclitaxel benefit.Trial registration numberUMIN Clinical Trials Registry: C000000082 (SAMIT); ClinicalTrials.gov identifier, 02628951 (South Korean trial)