A deep learning system for detection of early Barrett's neoplasia: a model development and validation study.

BACKGROUND: Computer-aided detection (CADe) systems could assist endoscopists in detecting early neoplasia in Barrett's oesophagus, which could be difficult to detect in endoscopic images. The aim of this study was to develop, test, and benchmark a CADe system for early neoplasia in Barrett's oesophagus. METHODS: The CADe system was first pretrained with ImageNet followed by domain-specific pretraining with GastroNet. We trained the CADe system on a dataset of 14 046 images (2506 patients) of confirmed Barrett's oesophagus neoplasia and non-dysplastic Barrett's oesophagus from 15 centres. Neoplasia was delineated by 14 Barrett's oesophagus experts for all datasets. We tested the performance of the CADe system on two independent test sets. The all-comers test set comprised 327 (73 patients) non-dysplastic Barrett's oesophagus images, 82 (46 patients) neoplastic images, 180 (66 of the same patients) non-dysplastic Barrett's oesophagus videos, and 71 (45 of the same patients) neoplastic videos. The benchmarking test set comprised 100 (50 patients) neoplastic images, 300 (125 patients) non-dysplastic images, 47 (47 of the same patients) neoplastic videos, and 141 (82 of the same patients) non-dysplastic videos, and was enriched with subtle neoplasia cases. The benchmarking test set was evaluated by 112 endoscopists from six countries (first without CADe and, after 6 weeks, with CADe) and by 28 external international Barrett's oesophagus experts. The primary outcome was the sensitivity of Barrett's neoplasia detection by general endoscopists without CADe assistance versus with CADe assistance on the benchmarking test set. We compared sensitivity using a mixed-effects logistic regression model with conditional odds ratios (ORs; likelihood profile 95% CIs). FINDINGS: Sensitivity for neoplasia detection among endoscopists increased from 74% to 88% with CADe assistance (OR 2·04; 95% CI 1·73-2·42; p<0·0001 for images and from 67% to 79% [2·35; 1·90-2·94; p<0·0001] for video) without compromising specificity (from 89% to 90% [1·07; 0·96-1·19; p=0·20] for images and from 96% to 94% [0·94; 0·79-1·11; ] for video; p=0·46). In the all-comers test set, CADe detected neoplastic lesions in 95% (88-98) of images and 97% (90-99) of videos. In the benchmarking test set, the CADe system was superior to endoscopists in detecting neoplasia (90% vs 74% [OR 3·75; 95% CI 1·93-8·05; p=0·0002] for images and 91% vs 67% [11·68; 3·85-47·53; p<0·0001] for video) and non-inferior to Barrett's oesophagus experts (90% vs 87% [OR 1·74; 95% CI 0·83-3·65] for images and 91% vs 86% [2·94; 0·99-11·40] for video). INTERPRETATION: CADe outperformed endoscopists in detecting Barrett's oesophagus neoplasia and, when used as an assistive tool, it improved their detection rate. CADe detected virtually all neoplasia in a test set of consecutive cases. FUNDING: Olympus.

Endoscopic Resection Without Subsequent Ablation Therapy for Early Barrett's Neoplasia: Endoscopic Findings and Long-Term Mortality.

INTRODUCTION: After endoscopic resection (ER) of neoplasia in Barrett's esophagus (BE), it is recommended to ablate the remaining BE to minimize the risk for metachronous disease. However, we report long-term outcomes for a nationwide cohort of all patients who did not undergo ablation of the remaining BE after ER for early BE neoplasia, due to clinical reasons or performance status. METHODS: Endoscopic therapy for BE neoplasia in the Netherlands is centralized in 8 expert centers with specifically trained endoscopists and pathologists. Uniformity is ensured by a joint protocol and regular group meetings. We report all patients who underwent ER for a neoplastic lesion between 2008 and 2018, without further ablation therapy. Outcomes include progression during endoscopic FU and all-cause mortality. RESULTS: Ninety-four patients were included with mean age 74 (± 10) years. ER was performed for low-grade dysplasia (LGD) (10%), high-grade dysplasia (HGD) (25%), or low-risk esophageal adenocarcinoma (EAC) (65%). No additional ablation was performed for several reasons; in 73 patients (78%), the main argument was expected limited life expectancy. Median C2M5 BE persisted after ER, and during median 21 months (IQR 11-51) with 4 endoscopies per patient, no patient progressed to advanced cancer. Seventeen patients (18%) developed HGD/EAC: all were curatively treated endoscopically. In total, 29/73 patients (40%) with expected limited life expectancy died due to unrelated causes during FU, none of EAC. CONCLUSION: In selected patients, ER monotherapy with endoscopic surveillance of the residual BE is a valid alternative to eradication therapy with ablation.

Investigating obscure gastrointestinal bleeding: capsule endoscopy or double balloon enteroscopy?

The possibility to visualise the small bowel has dramatically improved with the introduction of capsule endoscopy (CE) and double balloon enteroscopy (DBE). CE and DBE have become standard practice in investigating suspected diseases of the small bowel. An important reason to perform small bowel investigations is obscure gastrointestinal bleeding. To investigate obscure gastrointestinal bleeding, some advocate performing CE while others recommend DBE . In this systematic review, we provide an overview of studies in which patients with obscure gastrointestinal bleeding underwent both CE and DBE . These data show that CE and DBE have comparable diagnostic yields in the evaluation of obscure gastrointestinal bleeding of 50 to 60%. Therapeutic interventions using DBE were performed in 11 to 57% of cases. In most studies, there was good concordance between the two procedures but both techniques can be falsely negative. Given its safety, patient tolerability and ability to view the entire small bowel, CE can be recommended as the first investigation for obscure gastrointestinal bleeding, if necessary, followed by DBE . Finally, we provide an algorithm with practical guidelines for the evaluation ofobscure gastrointestinal bleeding.

First Report of Potato spindle tuber viroid in Cape Gooseberry (Physalis peruviana) from Turkey and Germany.

Since the recent identification of Potato spindle tuber viroid (PSTVd) in vegetatively propagated ornamental plant species (4), many growers have asked to have their mother plants tested for this viroid. In December of 2007, a grower from Turkey submitted cuttings of cape gooseberry (Physalis peruviana) to be tested for PSTVd. Initial testing by real-time reverse transcription (RT)-PCR according to Boonham et al. (1) indicated the presence of either Mexican papita viroid, PSTVd, or Tomato chlorotic dwarf viroid in four samples. To identify the viroid(s) present, isolated RNA from these samples was used for RT-PCR (2), and products of the expected full genome size for the three viroids were amplified from each sample. One of the PCR products was sequenced (GenBank Accession No. EU862230) and analysis of the 357 nt sequence indicated it was most related to PSTVd sequences belonging to the so-called 'Oceanian' strain of the viroid (3), with 99.7% identity to GenBank Accession No. AY962324. Therefore, the viroid was identified as PSTVd. Pathogenicity of this PSTVd genotype was demonstrated when 4 weeks after mechanical inoculation with sap extracts seedlings of tomato cv. Money-maker showed the expected viroid symptoms of chlorosis and stunting, and the presence of the viroid in these plants was confirmed by RT-PCR (2). In March of 2008, by use of RT-PCR (2) and sequencing of the PCR product (GenBank Accession No. EU862231), PSTVd was identified in young seedlings of P. peruviana from a German grower. The German isolate differed at only three nucleotide positions from the Turkish isolate. The identification of PSTVd in young seedlings indicates that seeds had been source of infection, whereas in the case of the PSTVd infected cuttings from Turkey, the infection originated from infected mother plants. To our knowledge, these are the first reports of PSTVd in P. peruviana. Although infected P. peruviana plants did not show symptoms, they might act as sources of inoculum for crops like potato and tomato, which may suffer serious damage. References: (1) N. Boonham et al. J. Virol. Methods 116:139, 2004. (2) A. M. Shamloul et al. Can. J. Plant Pathol. 19:89, 1997. (3) J. Th. J. Verhoeven et al. Eur. J. Plant Pathol. 110:823, 2004. (4) J. Th. J. Verhoeven et al. Plant Pathol. 57:399, 2008.

Correlation between predicted and actual consequences of capsule endoscopy on patient management.

BACKGROUND: Capsule endoscopy (CE) is a relatively new diagnostic modality in the evaluation of patients with suspected small bowel pathology. It is unclear to what extent physicians are able to predict the clinical consequences of CE on patient management. METHODS: In this prospective study, 180 consecutive CE examinations were analysed. Prior to CE, referring physicians were asked to indicate the consequences of CE according to potential different CE outcomes. The influence of CE on patient management was determined with at least 1 year follow-up. Management consequences were defined as major (surgical or endoscopic intervention, or medical therapy) or minor (nonspecific therapy, including iron supplementation, or no further diagnostic tests). RESULTS: CE led to major management consequences in 32% of cases. Of patients with obscure gastrointestinal bleeding and normal CE findings, 91% were independent of blood transfusions and experienced no further bleeding episodes during a mean follow-up of 33 months. In 78% of 118 cases that were evaluated, the actual consequences of CE matched the consequences predicted by the referring physicians. CONCLUSION: CE had a major impact on patient management in about one third of investigations. In the majority of cases, physicians adequately predicted the clinical consequences of CE.

Capsule endoscopy: a review from the clinician's perspectives.

The possibility of visualization of the small bowel has dramatically improved with the introduction of capsule endoscopy (CE) and it has rapidly become standard practice in investigating diseases of the small bowel. Nowadays, there are many reasons to perform CE: most often suspected obscure gastrointestinal bleeding, inflammatory bowel disease, celiac disease and small bowel neoplasia. CE has higher diagnostic yields than other (more invasive) diagnostic modalities for most of these indications. Furthermore, there are virtually no real contra-indications anymore for the procedure. CE appears also to be safe in selected paediatric cases in which it has similar diagnostic yields compared with adults. This review is a practical update for the clinician on CE, including the complementary role of double balloon enteroscopy and the new developments of Pillcam ESO and the Pillcam COLON.

Congenital malformations of the inner ear and the vestibulocochlear nerve in children with sensorineural hearing loss: evaluation with CT and MRI.

PURPOSE: The purpose of this work was to study the diagnostic value of CT and MRI in children with sensorineural hearing loss and to analyze anatomic abnormalities of the inner ear and the vestibulocochlear nerve in this patient group. METHOD: We evaluated 42 inner ears in 21 children with congenital deafness who had congenital inner ear malformations and who were candidates for cochlear implants. All patients were studied with high resolution MR and helical CT examinations. The MR study included a T2-weighted 3D fast SE sequence. We describe and tabulate the anatomic abnormalities. Special attention was given to abnormalities of the vestibulocochlear nerve. The field of view in the plane according to the length axis of the internal auditory canal (IAC) was 4 cm. Additional continuous parasagittal reformations perpendicular to the length axis of the IAC were studied with a field of view of 3 cm. RESULTS: CT and MRI allowed accurate identification of malformations of the inner ear in children with congenital deafness. We identified 99 malformations, with a majority of patients demonstrating multiple abnormalities. Common imaging findings were Mondini abnormality and Mondini variants (12/42) and fusion of the lateral or superior semicircular canal with the vestibule (12/42). MRI demonstrated in 9 of 21 patients a rudimentary or absent vestibulocochlear nerve in the auditory canal. CONCLUSION: CT and MRI are important modalities to analyze the inner ear in children who are candidates for cochlear implants. MRI with an extremely small field of view should be used to study possible abnormalities of the vestibulocochlear nerves. This may alter clinical care and allow cochlear implant placement in patients whose electrodiagnostic studies suggest that the implant should not be performed. The detailed analysis of abnormalities of the inner ear might establish prognostic factors.

Role of high magnification specimen radiography in surgical and core biopsies of the breast.

The clinical relevance of a high magnification specimen radiography (HMSR) system in breast biopsies was evaluated and compared with conventional specimen radiography with a mammography system (SRM). 100 surgical biopsies of 72 patients and 248 core biopsies of 30 patients were examined in (a) maximal 20-fold HMSR in combination with storage phosphors and (b) 1.8-fold SRM using a film-screen system. Detection of calcifications/soft tissue lesions and the impact on management were evaluated. In surgical biopsies, SRM could detect only 22% of individual microcalcifications, 39% of calcified lesions and 67% of soft tissue lesions identified with HMSR. Calcifications down to 10 microns were identified with HMSR. In five biopsies, peripheral calcifications leading to additional resection were recognized only with HMSR; in three of these they were indicative of malignant tissue. In core biopsies, only 12% of individual microcalcifications seen with HMSR were identified with SRM. 52% and 16% of all cores were calcified on HMSR and SRM, respectively. Microcalcifications within cores were found only with HMSR in 41% of patients with calcified lesions. In conclusion, the better detectability of microcalcifications with HMSR led to justified additional tissue resections in surgical patients and reduced the number of core biopsies required in interventional patients.

MRI of the breast in patients with metastatic disease of unknown primary.

The objective of this study was to assess the value of contrast-enhanced dynamic breast imaging in patients with carcinoma of unknown primary (CUP). Fourteen patients presenting with metastatic disease compatible with breast cancer (axillary lymph node metastasis: n = 6; supraclavicular lymph node metastasis: n = 1; bone metastasis: n = 3; liver metastasis: n = 3; lung metastasis: n = 1), who had no evidence of tumor in X-ray mammograms and ultrasound, underwent bilateral dynamic breast MR imaging. Suspicious lesions were localized preoperatively using a stereotactic device for MR-guided localization procedures. Magnetic resonance imaging revealed suspicious lesions in 9 of 14 patients. Histopathology revealed invasive carcinoma of the breast in 6 of these patients. Two enhancing lesions were fibroadenomas; one proved to be sclerosing adenosis. In 5 patients MR imaging showed no abnormality. Follow-ups performed up to 1 year after initial treatment revealed no breast cancers in these 5 patients. In patients with metastatic disease of unknown primary, MRI of the breast depicts the primary in a considerable number of cases with normal conventional evaluation.

MR imaging of mammographically detected clustered microcalcifications: is there any value?

PURPOSE: To assess the value of dynamic magnetic resonance (MR) imaging in patients with mammographically detected suspicious microcalcifications. MATERIALS AND METHODS: Sixty-three patients (age range, 31-84 years; mean age, 56 years) with mammographically suspicious clustered microcalcifications underwent dynamic MR imaging of the breast before surgical biopsy. MR imaging was performed with a 1.5-T unit and a two-dimensional fast low-angle shot sequence with a flip angle of 80 degrees. Thirty-two axial sections, 4 mm thick and without interval gaps, were acquired before and five times after administration of gadopentetate dimeglumine (dose, 0.1 mmol per kilogram of bodyweight). Histologic findings were used as the standard of reference. Any effect of MR imaging on surgical management was noted. RESULTS: Biopsy findings verified five patients with T1 invasive carcinomas, 33 with ductal carcinomas in situ, 13 with proliferative fibrocystic disease, eight with nonproliferative fibrocystic disease, and four with sclerosing adenosis. Contrast material-enhanced dynamic MR imaging had a sensitivity of 45%, a specificity of 72%, and an overall accuracy of 56% in differentiating benign from malignant microcalcifications. All invasive tumors were detected with MR imaging. In no case was surgical management altered by MR imaging findings. CONCLUSION: MR imaging of the breast is not reliable in differentiation of benign from malignant breast disease in mammographically suspicious clustered microcalcifications and has no effect on treatment.

[A new specimen radiography device with a maximum 20-fold magnification for breast diagnosis]. / Ein neues Präparatradiographiegerät mit maximal 20facher Vergrösserung für die Mammadiagnostik.

PURPOSE: A radiography system specially developed for specimen radiography and allowing maximal 20-fold magnification is presented. The efficiency of the system was tested and compared with that of conventional magnification mammography systems. METHODS: 23 surgical and 90 core biopsies of the breast were examined for detection of microcalcifications. As criteria the number of identifiable calcifications, their shape and configuration as well as tissue contrast were chosen. RESULTS: The new technique detected about 400% more microcalcifications, 200% more core and 50% more surgical biopsies containing calcifications. Thus, in a few cases, additional core biopsies were unnecessary. Moreover, this new system yielded additional information for the pathologist and surgeon concerning the exact localisation of suspicious lesions that facilitated working up specimens, or indicated additional surgical removal in special cases. CONCLUSIONS: By identification of malignant lesions not detectable with conventional magnification radiography systems, as well as a more exact localisation of suspicious lesions, false negative results may be reduced.

A comparison of conventional mammographic magnification, ultra high magnification and industrial magnification radiography in the radiographic detection of microcalcifications within core biopsies of the breast.

The objective was to compare conventional magnification radiography (CMR), ultra high magnification radiography (UHMR) and industrial magnification radiography (IMR) in the detection of microcalcifications in breast core biopsies. 440 core biopsies were examined in 1.8-fold CMR and in 7-fold UHMR using a prototype unit. A subgroup of 59 core biopsies were also examined in 10-fold IMR. Number, size, and demarcation of microcalcifications, as well as tissue contrast, were evaluated. Only 67% of the microcalcifications seen with UHMR were detected by CMR and 78% of the core biopsies showing calcifications in UHMR were calcified in CMR. Only 38% and 58% of microcalcifications verified by IMR were identified by CMR and UHMR, respectively. 47% and 63% of the core biopsies showing calcifications in IMR were calcified in CMR and UHMR, respectively. Tissue contrast of IMR was superior to both other modalities. On the other hand, increased cost and time will probably prohibit the use of IMR for specimen radiography in routine clinical examinations. In conclusion, UHMR identifies substantially more core biopsies with microcalcifications than CMR, thus potentially reducing the number of core biopsies needed for histological analysis. IMR allowed the detection of approximately 50%/160% more microcalcifications than UHMR/CMR, thus rendering it the reference mode.

Microcalcifications in breast core biopsy specimens: disappearance at radiography after storage in formaldehyde.

PURPOSE: To determine whether radiographically proved calcifications in core biopsy specimens are better preserved in a nonaqueous fixative than in aqueous solutions. MATERIALS AND METHODS: One hundred fifty core biopsy specimens were taken from two female breast specimens that had been preserved in ethanol. Forty-one of the specimens with radiographically proved microcalcifications were divided into four groups and deposited in four different solutions: 10% formaldehyde, 0.9% sodium chloride, electrolyte solution, and 74.1% ethanol with 10% 2-propanol. The core specimens were radiographed again after 1 and 3 days. Five ethanol-preserved specimens were also reexamined radiographically after 2 weeks. RESULTS: Within 3 days, total radiographic disappearance of microcalcifications was observed in all core biopsy specimens that were immersed in solutions with high water content. In those core specimens preserved in ethanol, microcalcifications showed no change. CONCLUSION: A nonaqueous fixative, such as ethanol, is a better preservative of microcalcifications in breast core biopsy specimens than various aqueous solutions, possibly because calcium compounds are water-soluble.

[Ductal carcinoma in situ in dynamic MR-mammography at 1.5 T]. / Das duktale In-situ-Karzinom in der dynamischen MR-Mammographie bei 1,5 T.

PURPOSE: To define the value of contrast-enhanced MR mammography in ductal carcinoma in situ (DCIS). MATERIAL AND METHODS: In a group of 35 patients with DCIS, the results of MR imaging were compared to histopathology and immunohistochemistry in a retrospective study. RESULTS: In 35 patients with DCIS, a signal enhancement was found in 25 cases (72%). In 15 of these cases, the signal time curve was typical for malignancy. The other 10 patients had non-specific signal curves. Six of 35 patients (11%) had no enhancement within the tumour region. Four of 35 patients (11%) had bilateral diffuse signal increase, and regions of DCIS could not be identified clearly. Three DCIS were visualised exclusively by MR mammography. The configuration of signal enhancement was sharp (32%), unsharp (48%) or dendritic (20%). DCIS of the comedo type showed a significantly higher enhancement than the non-comedo type. A significant correlation between the grade of vascularisation in immunohistochemistry and signal enhancement in MR mammography could not be demonstrated. CONCLUSION: Dynamic MR mammography does not reliably visualise DCIS.

Chromosomal studies after bone marrow transplantation for leukaemia in children.

Chromosomal studies were performed on 114 blood samples and 117 bone marrow samples, taken systematically over a period of 4 months after bone marrow transplantation (BMT) in 42 children grafted for acute lymphoblastic leukaemia (ALL) (n = 20), acute myeloid leukaemia (AML) (n = 16), non-Hodgkin's lymphoma (NHL) (n = 2) and myelodysplastic syndrome (MDS) (n = 4). In some cases, follow-up investigations were performed. In the first 4 months following BMT, mixed chimerism was frequently observed in blood of AML (25%), ALL (30%), NHL (100%) cases and in bone marrow samples of ALL (35%). The presence and relative number of a patient's own metaphases found shortly after transplantation was not related to leukaemia relapse and probably represents residual non-malignant haematopoietic precursor cells of the host. In only one child grafted for MDS with 45,XY, -7 karyotype was the marker clone still detectable in bone marrow at day +437 post-BMT. This patient shows no recurrence of the MDS and has a sustained haematological recovery at the time of writing, i.e. 4.5 years post-BMT. In 11 other patients, various structural chromosomal abnormalities (not related to the original leukaemia) were found in both peripheral blood and bone marrow. In three different patients structural anomalies were also found in bone marrow and blood samples from donor-derived cells. This indicates that, besides irradiation, there are other as yet unidentified factors (e.g. drugs), which are capable of inducing chromosomal anomalies in the post-BMT period.