A Swin Transformer and Residualnetwork Combined Model for Breast Cancer Disease Multi-Classification Using Histopathological Images

Breast cancer has become a killer of women's health nowadays.In order to exploit the potential representational capabilities of the models more comprehensively,we propose a multi-model fusion strategy.Specifically,we combine two differently structured deep learning models,ResNet101 and Swin Transformer(SwinT),with the addition of the Convolutional Block Attention Module(CBAM)attention mechanism,which makes full use of SwinT's global context information modeling ability and ResNet101's local feature extraction ability,and additionally the cross entropy loss function is replaced by the focus loss function to solve the problem of unbalanced allocation of breast cancer data sets.The multi-classification recognition accuracies of the proposed fusion model under 40X,100X,200X and 400X BreakHis datasets are 97.50%,96.60%,96.30 and 96.10%,respectively.Compared with a single SwinT model and ResNet 101 model,the fusion model has higher accuracy and better generalization ability,which provides a more effective method for screening,diagnosis and pathological classification of female breast cancer.

Using MsfNet to Predict the ISUP Grade of Renal Clear Cell Carcinoma in Digital Pathology Images

Clear cell renal cell carcinoma(ccRCC)represents the most frequent form of renal cell carcinoma(RCC),and accurate International Society of Urological Pathology(ISUP)grading is crucial for prognosis and treatment selection.This study presents a new deep network called Multi-scale Fusion Network(MsfNet),which aims to enhance the automatic ISUP grade of ccRCC with digital histopathology pathology images.The MsfNet overcomes the limitations of traditional ResNet50 by multi-scale information fusion and dynamic allocation of channel quantity.The model was trained and tested using 90 Hematoxylin and Eosin(H&E)stained whole slide images(WSIs),which were all cropped into 320×320-pixel patches at 40×magnification.MsfNet achieved a micro-averaged area under the curve(AUC)of 0.9807,a macro-averaged AUC of 0.9778 on the test dataset.The Gradient-weighted Class Activation Mapping(Grad-CAM)visually demonstrated MsfNet’s ability to distinguish and highlight abnormal areas more effectively than ResNet50.The t-Distributed Stochastic Neighbor Embedding(t-SNE)plot indicates our model can efficiently extract critical features from images,reducing the impact of noise and redundant information.The results suggest that MsfNet offers an accurate ISUP grade of ccRCC in digital images,emphasizing the potential of AI-assisted histopathological systems in clinical practice.

Research Progress on the Relationship between Magnetic Resonance Imaging and Clinical Pathological Characteristics of Liver Cancer

Liver cancer is one of the main malignant tumors in the digestive system.Early detection and treatment have positive significance in improving patient prognosis and reducing mortality.MRI is the main method for liver cancer examination,which mainly uses computers to compare imaging of different energy regions of tumors,observe the density and signal changes of liver cancer,and the degree of tumor enhancement.In particular,various new MRI functional imaging technologies,such as diffusion-weighted imaging,perfusion weighted imaging,delayed imaging,liver cell specific contrast agent enhanced imaging,etc.,can be used at the molecular level Multiple aspects such as cell function provide clinicians with richer diagnostic information.Therefore,further comparative analysis of MRI manifestations and pathological results of liver cancer can help to gain a deeper understanding of the biological behavior of tumors and provide a basis for treatment decision-making and prognosis evaluation.

Consistency between magnetic resonance diffusion-weighted images and pathological findings in a hyperacute cerebral infarction rabbit model

BACKGROUND： Because magnetic resonance diffusion-weighted imaging is sensitive to water molecule movement, it has particular advantages for early diagnosis of cerebral infarction. However, the relationship between apparent diffusion coefficient changes with ischemia time, particularly relative apparent diffusion coefficient and tissue pathological changes remains controversial. OBJECTIVE： To explore the correlation between apparent diffusion coefficient changes and pathologic changes in hyperacute cerebral infarction. DESIGN, TIME AND SETTING： A randomized, controlled, animal experiment of neuroimaging. The study was performed at the Laboratory of Radiology Department, Longgang Central Hospital of Shenzhen from October 2007 to October 2008. MATERIALS： Magnetic resonance scanner was purchased from Philips Medical Systems, Best, the Netherlands. METHODS： A total of 42 healthy, adult, New Zealand rabbits were randomly assigned into sham-operation, ischemia 0.5-, 1-, 2-, 3-, 4-, and 6-hour groups, with six animals in each group. Local cerebral ischemia model was established by right middle cerebral artery occlusion, and cranial MRI scanning and pathologic observation were performed, respectively, at 0.5, 1,2, 3, 4, and 6 hours following ischemia. The middle cerebral artery of sham-operation group was only exposed, but not occluded. Images at the above-mentioned time points were also collected. MAIN OUTCOME MEASURES： Apparent diffusion coefficient and relative apparent diffusion coefficient values of abnormal signal on diffusion-weighted imaging were calculated and compared with pathological changes in the ischemic region. RESULTS： No abnormal diffusion-weighted imaging signals or pathological changes were observed in the sham-operation group. Abnormal signal intensity on diffusion-weighted imaging was first observed in the 0.5-hour group. Apparent diffusion coefficient and relative apparent diffusion coefficient values decreased in all middle cerebral artery occlusion rabbits and reached lowest levels at 3 hours, followed by a gradual increase. The right ischemic basal ganglia region with high signal intensity on diffusion-weighted imaging extended with increasing time of occlusion, and the pathologic outcome corresponded with MRI changes. CONCLUSION： Relative apparent diffusion coefficient values changed regularly with ischemia time and displayed good correspondence to pathological manifestations.

A homological multi-information fusion method for processing gastric tumor tissue pathological images

A homological multi-information image fusion method was introduced for recognition of the gastric tumor pathological tissue images.The main purpose is that fewer procedures are used to provide more information and the result images could be easier to be understood than any other methods.First,multi-scale wavelet transform was used to extract edge feature,and then watershed morphology was used to form multi-threshold grayscale contours.The research laid emphasis upon the homological tissue image fusion based on extended Bayesian algorithm,which fusion result images of linear weighted algorithm was used to compare with the ones of extended Bayesian algorithm.The final fusion images are shown in Fig 5.The final image evaluation was made by information entropy,information correlativity and statistics methods.It is indicated that this method has more advantages for clinical application.

Factors associated with gastrointestinal stromal tumor rupture and pathological risk:A single-center retrospective study

BACKGROUND Gastrointestinal stromal tumor(GIST)is a rare gastrointestinal mesenchymal tumor with potential malignancy.Once the tumor ruptures,regardless of tumor size and mitotic number,it can be identified into a high-risk group.It is of great significance for the diagnosis,treatment,and prognosis of GIST if non-invasive examination can be performed before surgery to accurately assess the risk of tumor.AIM To identify the factors associated with GIST rupture and pathological risk.METHODS A cohort of 50 patients with GISTs,as confirmed by postoperative pathology,was selected from our hospital.Clinicopathological and computed tomography data of the patients were collected.Logistic regression analysis was used to evaluate factors associated with GIST rupture and pathological risk grade.RESULTS Pathological risk grade,tumor diameter,tumor morphology,internal necrosis,gas-liquid interface,and Ki-67 index exhibited significant associations with GIST rupture(P<0.05).Gender,tumor diameter,tumor rupture,and Ki-67 index were found to be correlated with pathological risk grade of GIST(P<0.05).Multifactorial logistic regression analysis revealed that male gender and tumor diameter≥10 cm were independent predictors of a high pathological risk grade of GIST[odds ratio(OR)=11.12,95%confidence interval(95%CI):1.81-68.52,P=0.01;OR=22.96,95%CI:2.19-240.93,P=0.01].Tumor diameter≥10 cm,irregular shape,internal necrosis,gas-liquid interface,and Ki-67 index≥10 were identified as independent predictors of a high risk of GIST rupture(OR=9.67,95%CI:2.15-43.56,P=0.01;OR=35.44,95%CI:4.01-313.38,P<0.01;OR=18.75,95%CI:3.40-103.34,P<0.01;OR=27.00,95%CI:3.10-235.02,P<0.01;OR=4.43,95%CI:1.10-17.92,P=0.04).CONCLUSION Tumor diameter,tumor morphology,internal necrosis,gas-liquid,and Ki-67 index are associated with GIST rupture,while gender and tumor diameter are linked to the pathological risk of GIST.These findings contribute to our understanding of GIST and may inform non-invasive examination strategies and risk assessment for this condition.

Comparison of imaging-based and pathological dimensions in pancreatic neuroendocrine tumors

AIM To establish the ability of magnetic resonance(MR) and computer tomography(CT) to predict pathologic dimensions of pancreatic neuroendocrine tumors(Pan NET) in a caseload of a tertiary referral center.METHODS Patients submitted to surgery for Pan NET at the Surgical Unit of the Pancreas Institute with at least 1 preoperative imaging examination(MR or CT scan) from January 2005 to December 2015 were included and data retrospectively collected. Exclusion criteria were: multifocal lesions, genetic syndromes, microadenomas or mixed tumors, metastatic disease and neoadjuvant therapy. Bland-Altman(BA) and Mountain-Plot(MP) statistics were used to compare size measured by each modality with the pathology size. Passing-Bablok(PB) regression analysis was used to check the agreement between MR and CT.RESULTS Our study population consisted of 292 patients. Seventy-nine(27.1%) were functioning Pan NET. The mean biases were 0.17 ± 7.99 mm, 1 ± 8.51 mm and 0.23 ± 9 mm, 1.2 ± 9.8 mm for MR and CT, considering the overall population and the subgroup of non-functioning-Pan NET, respectively. Limits of agreement(LOA) included the vast majority of observations, indicating a good agreement between imaging and pathology. The MP further confirmed this finding and showed that the two methods are unbiased with respect to each other. Considering ≤ 2 cm non-functioning-Pan NET, no statistical significance was found in the size estimation rate of MR and CT(P = 0.433). PBR analysis did not reveal significant differences between MR, CT and pathology.CONCLUSION MR and CT scan are accurate and interchangeable imaging techniques in predicting pathologic dimensions of Pan NET.

The Epidemiological, Clinical and Radiological (Echography, Mammography) Characteristics of Breast Pathologies at the Diagnostic Imaging Center (C.I.D) “Teriya” in Bamako

Breast pathology is varied, bringing together tumor and non-tumor lesions. Objective: To study the contribution of the ultrasound-mammography pair in the diagnosis of breast pathologies. Materials and Method: This was a retrospective descriptive study, carried out over a period of 3 years (from January 2018 to December 2020) at the Diagnostic Imaging Center (C.I.D) “TERIYA” in BAMAKO. It concerned all patients who came for a mammogram/ultrasound examination of the breast. All women admitted for mammogram or breast ultrasound who were diagnosed with a breast injury during the study period were included. Incomplete records and radiological checks were not included. The variables analyzed were age, sex, clinical data, and ultrasound and mammography aspects. The devices used are: a Voluson 730 PRO ultrasound machine and a G 600T type mammography machine. Results: At the end of our study, we collected 254 breast pathologies on a number of 382 women, i.e. a frequency of 66.49%. The average age of our patients was 41 years old. The dominant clinical data were mastodynia (41.88%) and mammary nodule (21.70%). On imaging (mammo-ultrasound) the lesions predominated on the left in 36% of cases, bilateral in 28% of cases and in the upper-outer quadrants in 31.5% of cases. Tumor pathologies represented 66.54% of which 45.27% were benign mainly composed of fibro-adenoma (20.88%) and cyst (18.50%), 11.8% of suspected cases and 9.45% of cancers. Non-tumor pathologies represented 33.46%, mainly mastitis (16.14%), galactophoric dilations (11.02%) and abscesses (5.51%). These pathologies were classified in 50.3% in ACR2, 17.75% in ACR3 and 4, and in 14.20% in ACR5. Lymphadenopathy was present in 73.21% of cases.

What benefit can be obtained from magnetic resonance imaging diagnosis with artificial intelligence in prostate cancer compared with clinical assessments?

The present study aimed to explore the potential of artificial intelligence(AI)methodology based on magnetic resonance(MR)images to aid in the management of prostate cancer(PCa).To this end,we reviewed and summarized the studies comparing the diagnostic and predictive performance for PCa between AI and common clinical assessment methods based on MR images and/or clinical characteristics,thereby investigating whether AI methods are generally superior to common clinical assessment methods for the diagnosis and prediction fields of PCa.First,we found that,in the included studies of the present study,AI methods were generally equal to or better than the clinical assessment methods for the risk assessment of PCa,such as risk stratification of prostate lesions and the prediction of therapeutic outcomes or PCa progression.In particular,for the diagnosis of clinically significant PCa,the AI methods achieved a higher summary receiver operator characteristic curve(SROC-AUC)than that of the clinical assessment methods(0.87 vs.0.82).For the prediction of adverse pathology,the AI methods also achieved a higher SROC-AUC than that of the clinical assessment methods(0.86 vs.0.75).Second,as revealed by the radiomics quality score(RQS),the studies included in the present study presented a relatively high total average RQS of 15.2(11.0–20.0).Further,the scores of the individual RQS elements implied that the AI models in these studies were constructed with relatively perfect and standard radiomics processes,but the exact generalizability and clinical practicality of the AI models should be further validated using higher levels of evidence,such as prospective studies and open-testing datasets.

Comparative Study of Images with Pathology:Superparamagnetic Iron Oxide-enhanced Magnetic Resonance Image(MRI)of Splenic VX2 Tumor in Rats

Objective： To establish a rodent model of VX2 tumor of the spleen, to analyze relationship between the change of the signal intensity on superparamagnetic iron oxide enhanced magnetic resonance image （MRI） and pathologic change to evaluate the ability of superparamagnetic iron oxide enhanced MRI for detection of splenic metastases. Methods： 8 rodent models of VX2 tumor of spleen were established successfully. The images were obtained before and after administration of superparamagnetic iron oxide. T1-weighted spin-echo （SE） pulse sequence with a repetition time （TR） of 450 msec, and echo time （TE） of 12 msec （TR/TE=450/12） was used. The imaging parameters of T2-weighted SE pulse sequence were as follows： TR/TE=4000/128. Results： On plain MR scanning T1-weighted splenic VX2 tumor showed hypointensity or isointensity which approximated to the SI of splenic parenchyma. Therefore all lesions were not displayed clearly. On superparamagnetic iron oxide enhancement T2WI sequence the SI of splenic parenchyma decreased obviously with percentage of signal intensity loss （PSIL） of 55.04%, But the SI of tumor was not evidently changed with PSIL of 0.87%. Nevertheless the SNR of normal splenic parenchyma around the lesions had obvious difference （P〈0.001） comparatively. Therefore the contrast between tumor and spleen increased, and tumor displayed more clearly. Moreover the contrast-to-noise （CNR） between VX2 tumor and splenic parenchyma had an evident difference before and after admininstration of superparamagnetic iron oxide （P〈0.001）. Conclusion： On superparamagnetic iron oxide enhancement T1WI sequence the contrast of tumor-to-spleen is poor. Therefore it is not sensitive to characterize the lesions in spleen. On superparamagnetic iron oxide enhanced T2WI the contrast degree of lesions increases obviously. Consequently, superparamagnetic iron oxide -enhanced T2WI MRI scanning can improve the rate of detection and characterization for lesions of spleen.

Contrast investigation of multi-slice spiral CT perfusion imaging and pathological findings in VX2 soft-tissue tumor of rabbits

Objective: To perform a contrast investigation of multi-slice spiral CT (MSCT) perfusion imaging and pathological findings in VX2 soft-tissue tumor of rabbits, and discuss the applicative value of multi-slice spiral CT perfusion imaging in soft-tissue tumors. Methods: 8 Newzealand white rabbits were implanted with 0.1 ml VX2 tumor tissue suspension in bilateral proximal thighs. 14 days and 21 days later, CT plain scan and perfusion scan were performed on these rabbits respectively, then the images were transmitted to AW4.0 workstation, the functional maps and perfusion parameters including blood flow (BF), blood volume (BV), mean transit time (MTT) and permeability surface (PS) were computed and analyzed. Subsequently, the rabbits were sacrificed, the tumors of which were taken out for pathological examination. The correlation between MSCT functional parametric images and pathological findings was analyzed. Results: All the values of BF, BV and PS of VX2 soft-tissue tumors were obviously higher while the MTT-values were lower than those of the normal muscular tissues significantly (P<0.001). Conclusion: Multi-slice spiral CT perfusion imaging is an accurate, convenient and relatively safe functional imaging technique, which can give a quantitative assessment to blood perfusion and angiogenesis of soft-tissue tumors.

The Comparison of the Manifestation of the Clinical Imageology and Pathology between the Brucellar Spondylitis and the Spine Turberculosis

Objective: To improve the clinical differential diagnosis level, the clinical manifestation of the brucellar spondylitis and the spine turberculosis were discussed in this paper. Method: The study was completed in the No. 1 Affiliated Hospital of Hebei North University in Zhangjiakou City, Hebei Province, China, from January 2001 to December 2013 by Analyzing the X-ray, CT scanning and MRI of 257 cases of the brucellar spondylitis retrospectively and comparing with the clinical imageology and pathology 332 cases of turberculosis of the spine diagnosed finally. Results: The brucellar spondylitis: The focuses usually locate in the lumbar vertebra and L4, 5 has the highest occurrence rate. The focuses are often small but multiple, and limited to the edge of the vertebra. Hyperostosis and osteoscterosis are usually found in the tissuses around the focuses. There are often new focuses in the newborn bones, and the destruction of intervertebral discs is usually slight. Hyperostosis and osteoscterosis might be found in the surfaces of the joints. The densites of the bones close to the focuses become high. There were less or no paravertebral abscesses but inflammational granuloma can be found frequently. Turberculosis of the spine: The focuses are usually located in the thoracic and lumbar vertebra, and are characterized by the destruction of the vertebra and the intervertebral discs, accompanied by the appearance of dead bones. In most cases, paravertebral abscesses and osteoporosis might be found. Conclusions: The specific manifestation of the clinical imageology can help to differentiate the brucelar spondylitis from the turberculosis of the spine.

Does diffusion tensor data reflect pathological changes in the spinal cord with chronic injury?

Magnetic resonance diffusion tensor imaging has been shown to quantitatively measure the early pathological changes in chronic cervical spondylotic myelopathy. In this study, a novel spongy poly- urethane material was implanted in the rat C^-s epidural space to establish a rat model of chronic cervical spondylotic myelopathy. Diffusion tensor data were used to predict pathological changes. Results revealed that the fractional anisotropy value gradually decreased at 4, 24, and 72 hours and 1 week after injury in rat spinal cord, showing a time-dependent manner. Average diffusion coeffi- cient increased at 72 hours and 1 week after implantation. Hematoxylin-eosin staining and Luxol-fast-blue staining exhibited that the number of neurons in the anterior horn of the spinal cord gray matter and the nerve fiber density of the white matter gradually reduced with prolonged com- pression time. Neuronal loss was most significant at 1 week after injury. Results verified that the fractional anisotropy value and average diffusion coefficient reflected the degree of pathological change in the site of compression in rat models at various time points after chronic spinal cord compression injury, which potentially has a reference value in the early diagnosis of chronic cervical spondylotic myelopathy.

MRI Features of Intracranial Primitive Neuroectodermal Tumors in Adults:Comparing with Histopathological Findings

The MRI appearances of 7 adult patients with pathologically proven intracranial primitive neuroectodermal tumors (PNET) were retrospectively analyzed. The MRI features were compared with findings in pathology and surgery. In this group, the tumor masses were most commonly found in the semisphere of cerebrum and in the vermis of cerebellum. They were relatively large and 4 were in lobulated shape. All of them had well defined margins. MR images showed the tumors to be mildly or obviously hypointense on T 1 weighted images and hyperintense on T 2 weighted images. Most masses had heterogeneous appearances with some cystic and necrotic areas. Intratumoral haemorrhage and focal calcification were occasionally seen. Mostly, there was no or only mild surrounding edema. Marked inhomogeneous contrast enhancement on MRI was seen in 6 cases except one. Two patients with multiple intracranial metastases were revealed on MR images. In this series, there was good correlation between MRI features and findings in pathology and surgery. These results showed that certain MRI features might suggest the diagnosis of intracranial PNET in adults. MRI is an effective technique to detect these tumors and is helpful to treatment planning and follow up.

Imaging of Degenerative Pathologies of the Lumbar Spine: Professional Habits in Some Hospitals in Cameroon

Background: The lumbar spine is the portion most frequently involved in degenerative pathologies. Everyone will suffer one day from “low back pain”. These pathologies are very frequent: epidemiological studies have shown that 65% to 90% of the general population could be affected by low back pain (lumbago) which could become chronic at acertain stage or could be complicated (2% to 4%) of cases could end up affecting nerve roots. Chronic low back pain causes a major public health problem in terms of morbidity and socioeconomic repercussions. Purpose: The overall purpose of this study is to evaluate the professional habits of exam applicants in case of degenerative pathologies of the lumbar spine, to Appreciate the knowledge of clinicians on the usefulness of medical imaging techniques in the event of suspicion of a degenerative pathology of the lumbar spine and finally to evaluate if the professional habits of requesting examinations in Cameroon comply with the Recommendations for Clinical Practice. Methods: A cross-sectional and descriptive study was used and was based on questionnaires distributed to those practitioners who requested for these diagnostic medical imaging studies and procedures at Yaoundé General Hospital (HGY), Yaoundé Central Hospital, La Cathédrale Medical Center, Yaoundé University Teaching Hospital, the Douala General Hospital, the Laquintinie Hospital of Douala, the Military Hospital of Douala, the Daniel Muna Memorial Clinic of Douala, the International Center for Clinical and Medicine Imaging, the Ngaoundere Regional Hospital and the Protestant Hospital of Ngaoundere (HPN), from April 2020 to March 2021. Data collected was processed and analyzed via Epi Info version 12.0 and the statistical test used for correlation was Chi2. Results: 137 practitioners were retained among whom, 90 were male and 47 females, their average age being 46 years with working experience less than 5 years (35.8%). The results obtained show that parameters such as availability and accessibility of the required diagnostic medical imaging modalities could greatly influence the examination prescription. On the contrary, few prescribers were less interested in the costs and secondary effects due to irradiation and the invasiveness of these examinations. Professional habits of those who requested these examinations were in majority closer in line with the recommended clinical practices. However, an average of 43.78% of prescribers never followed recommendations nor applied them. On the other hand, knowledge and the application of these recommendations increased as the prescriber’s working experience too increased. Up to 54.47% of prescribers were not aware of those recommendations about requesting these examinations. Conclusion: Our results indicate that Medical Diagnostic Imaging techniques are not judiciously and optimally exploited in the diagnosis of degenerative pathologies of the lumbar spine and it would be necessary to establish recommendations for clinical practice adapted to Cameroonian realities.

Pathological tissue segmentation by using deep neural networks

Objective:The process of manually recognize the lesion tissue in pathological images is a key,laborious and subjective step in tumor diagnosis.An automatic segmentation method is proposed to segment lesion tissue in pathological images.Methods:We present a region of interest(ROI)method to generate a new pre-training dataset for training initial weights on DNNs to solve the overfitting problem.To improve the segmentation performance,a multiscale and multi-resolution ensemble strategy is proposed.Our methods are validated on a public segmentation dataset of colonoscopy images.Results:By using the ROI pre-training method,the Dice score of DeepLabV3 and ResUNet increases from 0.607 to 0.739 and from 0.572 to 0.741,respectively.The ensemble method is used in the testing phase,the Dice score of DeepLabV3 and ResUNet increased to 0.760 and 0.786.Conclusion:The ROI pre-training method and ensemble strategy can be applied to DeepLabV3 and ResUNet to improve the segmentation performance of colonoscopy images.

The Magnetic Resonance Image and Pathology of Spinal Cord Cavernous Hemangioma

Objective To investigate the characteristics of magnetic resonance image (MRI) of spinal cord cavernous hemangioma. Methods Six cases of spinal cord cavernous heman-gioma diagnosed by MRI and confirmed by pathology were reviewed. The characteristics of MRI were analyzed and correlated with pathological characteristics of spinal cord cavernous hemangioma. Results In 4 cases, the tumors were located in thoracic segment of the spinal cord and 2 in cervical cord. All lesions were solitary and the spinal cords were normal or a little thicker. The MRI showed that the images of focus were ball-like popcorn or mulberry with mixed signal,with short T2 signal around the focus. Under microscope, the hemangioma was composed of highly expanded blood sinusoids and its wall was thin and consisted of flat epithelial cells. There were some red blood cells in the cavity of the sinusoid and a little fibrous tissue in the diazoma between blood sinusoids. And also some fresh and old hemorrhages could be seen in the specimen. Conclusion MRI has high sensitivity and specificity in the diagnosis of spinal cavernous hemangioma.

Impact of blurs on machine-learning aided digital pathology image analysis

BACKGROUND Digital pathology image(DPI)analysis has been developed by machine learning(ML)techniques.However,little attention has been paid to the reproducibility of ML-based histological classification in heterochronously obtained DPIs of the same hematoxylin and eosin(HE)slide.AIM To elucidate the frequency and preventable causes of discordant classification results of DPI analysis using ML for the heterochronously obtained DPIs.METHODS We created paired DPIs by scanning 298 HE stained slides containing 584 tissues twice with a virtual slide scanner.The paired DPIs were analyzed by our MLaided classification model.We defined non-flipped and flipped groups as the paired DPIs with concordant and discordant classification results,respectively.We compared differences in color and blur between the non-flipped and flipped groups by L1-norm and a blur index,respectively.RESULTS We observed discordant classification results in 23.1%of the paired DPIs obtained by two independent scans of the same microscope slide.We detected no significant difference in the L1-norm of each color channel between the two groups;however,the flipped group showed a significantly higher blur index than the non-flipped group.CONCLUSION Our results suggest that differences in the blur-not the color-of the paired DPIs may cause discordant classification results.An ML-aided classification model for DPI should be tested for this potential cause of the reduced reproducibility of the model.In a future study,a slide scanner and/or a preprocessing method of minimizing DPI blur should be developed.

Artificial intelligence in pathological evaluation of gastrointestinal cancers

The integration of artificial intelligence(AI)has shown promising benefits in many fields of diagnostic histopathology,including for gastrointestinal cancers(GCs),such as tumor identification,classification,and prognosis prediction.In parallel,recent evidence suggests that AI may help reduce the workload in gastrointestinal pathology by automatically detecting tumor tissues and evaluating prognostic parameters.In addition,AI seems to be an attractive tool for biomarker/genetic alteration prediction in GC,as it can contain a massive amount of information from visual data that is complex and partially understandable by pathologists.From this point of view,it is suggested that advances in AI could lead to revolutionary changes in many fields of pathology.Unfortunately,these findings do not exclude the possibility that there are still many hurdles to overcome before AI applications can be safely and effectively applied in actual pathology practice.These include a broad spectrum of challenges from needs identification to cost-effectiveness.Therefore,unlike other disciplines of medicine,no histopathology-based AI application,including in GC,has ever been approved either by a regulatory authority or approved for public reimbursement.The purpose of this review is to present data related to the applications of AI in pathology practice in GC and present the challenges that need to be overcome for their implementation.

Value of MRI diffusion weighted imaging in localization of prostate cancer with whole-mount step section pathology

Objective To evaluate the value of MRI diffusion weighted imaging in localization of prostate cancer with whole-mount step section pathology. Methods We treated 36 patients using laparoscopic radical prostatectomy from Oct. 2009 to Jun. 2010. Patients who did not have an MRL /DWI examination or a surgical history of pros-