Attending Doctor Research Articles

Impact of Deep Learning-Based Computer-Aided Detection and Electronic Notification System for Pneumothorax on Time to Treatment: Clinical Implementation

ObjectiveTo assess whether the implementation of deep learning (DL) computer-aided detection (CAD) that screens for suspected pneumothorax (PTX) on chest radiography (CXR) combined with an electronic notification system (ENS) that simultaneously alerts both the radiologist and the referring clinician would affect time to treatment (TTT) in a real-world clinical practice. MethodsIn May 2022, a commercial deep learning-based CAD and ENS (DL-CAD-ENS) was introduced for all CXRs at an 818-bed general hospital, with 33 attending doctors and their residents using ENS, while 155 others used only CAD. We used difference-in-differences estimates to compare TTT between the CAD and ENS group and the CAD-only group for the period from January 2018 to April 2022 and from May 2022 to April 2023. ResultsA total of 603,028 CXRs from 140,841 unique patients were included, with a PTX prevalence of 2.0%. There was a significant reduction in TTT for supplemental oxygen therapy for the CAD and ENS group compared with the CAD-only group in the post-implementation period (-143.8 min; 95% confidence interval [CI], -277.8 to -9.9; P = .035). However, there was no significant difference in TTT for other treatments, including aspiration or tube-thoracostomy (14.4 min; 95% CI, -35.0 to 63.9) and consultation with the thoracic and cardiovascular surgery department (86.3 min; 95% CI, -175.1 to 347.6). ConclusionThe introduction of a DL-CAD-ENS reduced the time to initiate oxygen supplementation for patients with PTX.

What do patients with heart failure disclose about medication adherence at home to their hospital and primary care doctors? Exploratory interaction-based observational cohort study

ObjectivesThe main objective of this study was twofold: to investigate what kind of information patients with heart failure (HF) tell their doctors about their medication adherence at home, and how...

Patients with Hypochondriasis (Illness Anxiety Disorder): Key Elements of Effective Communication in the Dental Practice

The world data show a large number of people with illness anxiety disorder (hypochondriasis) among the population all over the world. Russia is no exception: hypochondriacal patients often seek dental care. A dentist should be able to identify this category of patients. As a rule, these patients "know everything": their diagnosis and the most modern and effective methods of treatment. They are truly literate people. Their medical literacy is the result of watching popular shows, reading articles in trendy magazines and medical websites. Frequently, these patients are not emotionally stable; they have been stressed for a long time, or are undergoing long-term hormone therapy. Identifying these patients is the key to the successful practice. This article describes an interesting clinical case of the patient, who was treated for 2 months by 5 different specialists and took strong medications. During treatment her condition only worsened: allergic reaction developed to medications, with lesion of mucous membrane of oral cavity. Her reluctance to stop taking medications, insistence on consulting a dentist-oncologist alerted the attending doctor. Analysis of medical history revealed that the patient had experienced severe stress in the recent past and was currently undergoing a course of hormonal corrective therapy. A clinical psychologist confirmed the dentist's suggestion that the patient had hypochondriacal disorder. Patients with hypochondria - who are they? How to recognize them? If you meet such patients at a dental appointment, how you should communicate with them? How to treat them? A good result is achieved by establishing contact with the patients and their family and friends. Sometimes it is impossible to do without the help of a clinical psychologist. Only this specialist can definitively decide: does the patient have hypochondriacal disorder or not? If necessary, the clinical psychologist provides special treatment. The probability of a favorable outcome with this approach is high.

Gap analysis between trainees' subjective competencies and the competencies expected by instructors in urology: A need assessment survey in Japan.

According to the rapid progress in surgical techniques, a growing number of procedures should be learned during postgraduate training periods. This study aimed to clarify the current situation regarding urological surgical training and identify the perception gap between trainees' competency and the competency expected by instructors in Japan. Regarding the 40 urological surgical procedures selected via the Delphi method, we collected data on previous caseloads, current subjective autonomy, and confidence for future skill acquisition from trainees (<15 post-graduate years [PGY]), and the competencies when trainees became attending doctors expected by instructors (>15 PGY), according to a 5-point Likert scale. In total, 174 urologists in Hokkaido Prefecture, Japan were enrolled in this study. The response rate was 96% (165/174). In a large proportion of the procedures, caseloads grew with accumulation of years of clinical practice. However, trainees had limited caseloads of robotic and reconstructive surgeries even after 15 PGY. Trainees showed low subjective competencies at present and low confidence for future skill acquisition in several procedures, such as open cystectomy, ureteroureterostomy, and ureterocystostomy, while instructors expected trainees to be able to perform these procedures independently when they became attending doctors. Trainees showed low subjective competencies and low confidence for future skill acquisition in several open and reconstructive procedures, while instructors considered that these procedures should be independently performable by attending doctors. We believe that knowledge of these perception gaps is helpful to develop a practical training program.

Treatment Preferences and Factors Influencing the Management of Primary Angle-Closure Suspect in China: A National Survey Study.

This study explored the current views of ophthalmologists in China on the preferred strategy and the primary influencing factors in managing primary angle-closure suspect (PACS). An online survey was distributed via WeChat to ophthalmologists working in hospitals at all levels throughout China to investigate the optimal management strategy and factors influencing the strategy for treating PACS from February to March 2023. Management strategies examined included regular observation (RO), laser peripheral iridectomy (LPI), and phacoemulsification with intraocular lens implantation (PEI). The study explored the factors that influenced clinical decision-making processes and collected demographic information from the respondents for analysis. A total of 1009 ophthalmologists responded; 442 responders considered RO as a reasonable option for patients with PACS, while the majority opted for more positive treatments, with 460 choosing LPI and 107 selecting PEI. Further grouping analysis revealed significant differences between tertiary and lower-level hospitals (P < 0.01) and between public and private hospitals (P < 0.05). Significant differences were found among resident, attending, and senior doctors and between northern and southern hospitals. The top three considerations for ophthalmologists who selected RO were ultrasound biomicroscopy (UBM) without pupillary block, absence of a family history of glaucoma, and good compliance with regular follow-up. For LPI, the factors were UBM indicating pupillary block, glaucoma symptoms, and poor compliance with regular follow-up. For PEI the factors were age, family history of glaucoma, and poor compliance with regular follow-up. The majority of ophthalmologists prefer aggressive treatments such as LPI or PEI over RO when dealing with patients with PACS. Disparity existed among ophthalmologists based on their level of experience, hospital ownership and level, and geographical location. The survey underscores the complexity and multifactorial nature of PACS management and highlights the need for further research and standardization of treatment protocols to ensure optimal patient outcomes.

Dyspnoea in liver cirrhosis: beyond the ordinary cardiopulmonary cause

BackgroundDyspnoea is a common presentation to the casualties, attributed mostly to cardiopulmonary diseases. Approach to dyspnoea consists of initial acute management to ensure airway patency, respiratory support, and hemodynamic stability, followed by identification of the underlying disease for a definite management to reduce short-term inpatient morbidity and mortality, and rehospitalization in long term. Awareness of the patient’s pre-existing medical conditions is a crucial clue to guide the attending doctor to the primary cause. Dyspnea in chronic liver disease can be due to concomitant cardiopulmonary diseases as well as pulmonary vascular congestion or abnormality associated with portal hypertension such as hepatopulmonary syndrome.Case presentationA 59-year-old man presented with a syncopal attack associated with breathlessness after physical exertion. He had alcohol associated liver cirrhosis for 2 years with a 6-month duration of platypnea and reduced effort tolerance. His physical examination revealed a cyanosed man with normal blood pressure and heart rate, without evidence of systemic and pulmonary fluid congestions. Chest radiography and computed tomography showed no evidence of pulmonary embolism and parenchymal disease and a normal echocardiography ruled out any cardiogenic cause. A diagnosis of hepatopulmonary syndrome was made after the appearance of microbubbles over left atrium at fourth to fifth cardiac cycle demonstrated by a transthoracic contrast-enhanced echocardiography to suggest an extracardiac intrapulmonary shunt.ConclusionsHPS is not an uncommon complication of chronic liver disease. Usual manifestation of HPS is platypnea which needs to be differentiated from orthopnoea associated with cardiopulmonary disease. Nevertheless, there is no effective therapeutic approach of HPS besides liver transplantation.

Visual Cascaded-Progressive Convolutional Neural Network (C-PCNN) for Diagnosis of Meniscus Injury.

The objective of this study is to develop a novel automatic convolutional neural network (CNN) that aids in the diagnosis of meniscus injury, while enabling the visualization of lesion characteristics. This will improve the accuracy and reduce diagnosis times. We presented a cascaded-progressive convolutional neural network (C-PCNN) method for diagnosing meniscus injuries using magnetic resonance imaging (MRI). A total of 1396 images collected in the hospital were used for training and testing. The method used for training and testing was 5-fold cross validation. Using intraoperative arthroscopic diagnosis and MRI diagnosis as criteria, the C-PCNN was evaluated based on accuracy, sensitivity, specificity, receiver operating characteristic (ROC), and evaluation performance. At the same time, the diagnostic accuracy of doctors with the assistance of cascade- progressive convolutional neural networks was evaluated. The diagnostic accuracy of a C-PCNN assistant with an attending doctor and chief doctor was compared to evaluate the clinical significance. C-PCNN showed 85.6% accuracy in diagnosing and identifying anterior horn injury, and 92% accuracy in diagnosing and identifying posterior horn injury. The average accuracy of C-PCNN was 89.8%, AUC = 0.86. The diagnosis accuracy of the attending physician with the aid of the C-PCNN was comparable to that of the chief physician. The C-PCNN-based MRI technique for diagnosing knee meniscus injuries has significant practical value in clinical practice. With a high rate of accuracy, clinical auxiliary physicians can increase the speed and accuracy of diagnosis and decrease the number of incorrect diagnoses.

Metal artifact reduction and clinical verification in oral and maxillofacial region based on deep learning

Objective: To construct a kind of neural network for eliminating the metal artifacts in CT images by training the generative adversarial networks (GAN) model, so as to provide reference for clinical practice. Methods: The CT data of patients treated in the Department of Radiology, West China Hospital of Stomatology, Sichuan University from January 2017 to June 2022 were collected. A total of 1 000 cases of artifact-free CT data and 620 cases of metal artifact CT data were obtained, including 5 types of metal restorative materials, namely, fillings, crowns, titanium plates and screws, orthodontic brackets and metal foreign bodies. Four hundred metal artifact CT data and 1 000 artifact-free CT data were utilized for simulation synthesis, and 1 000 pairs of simulated artifacts and metal images and simulated metal images (200 pairs of each type) were constructed. Under the condition that the data of the five metal artifacts were equal, the entire data set was randomly (computer random) divided into a training set (800 pairs) and a test set (200 pairs). The former was used to train the GAN model, and the latter was used to evaluate the performance of the GAN model. The test set was evaluated quantitatively and the quantitative indexes were root-mean-square error (RMSE) and structural similarity index measure (SSIM). The trained GAN model was employed to eliminate the metal artifacts from the CT data of the remaining 220 clinical cases of metal artifact CT data, and the elimination results were evaluated by two senior attending doctors using the modified LiKert scale. Results: The RMSE values for artifact elimination of fillings, crowns, titanium plates and screws, orthodontic brackets and metal foreign bodies in test set were 0.018±0.004, 0.023±0.007, 0.015±0.003, 0.019±0.004, 0.024±0.008, respectively (F=1.29, P=0.274). The SSIM values were 0.963±0.023, 0.961±0.023, 0.965±0.013, 0.958±0.022, 0.957±0.026, respectively (F=2.22, P=0.069). The intra-group correlation coefficient of 2 evaluators was 0.972. For 220 clinical cases, the overall score of the modified LiKert scale was (3.73±1.13), indicating a satisfactory performance. The scores of modified LiKert scale for fillings, crowns, titanium plates and screws, orthodontic brackets and metal foreign bodies were (3.68±1.13), (3.67±1.16), (3.97±1.03), (3.83±1.14), (3.33±1.12), respectively (F=1.44, P=0.145). Conclusions: The metal artifact reduction GAN model constructed in this study can effectively remove the interference of metal artifacts and improve the image quality.

Use of a Convolutional Neural Network to Predict the Malignant Potential of Gastrointestinal Stromal Tumors in Transabdominal Ultrasound Images: Visualization of the Focus of the Prediction Model

We established a deep convolutional neural network (CNN) model based on ultrasound images (US-CNN) for predicting the malignant potential of gastrointestinal stromal tumors (GISTs). A total of 980 ultrasound images from 245 pathology-confirmed GIST patients after surgical operation were retrospectively collected and divided into a low (very-low-risk, low-risk) and a high (medium-risk, high-risk) malignant potential group. Eight pre-trained CNN models were used to extract the features. The CNN model with the highest accuracy in the test set was selected. The model's performance was evaluated by calculating accuracy, sensitivity, specificity, positive-predictive value (PPV), negative-predictive value (NPV) and the F1 score. Three radiologists with different experience levels also predicted the malignant potential of GISTs in the same test set. US-CNN and human assessments were compared. Subsequently, gradient-weighted class activation diagrams (Grad-CAMs) were used to visualize the model's final classification decisions. Among the eight transfer learning-based CNNs, ResNet18 performed best. The accuracy, sensitivity, specificity, PPV, NPV and F1 score were 0.88, 0.86, 0.89, 0.82, 0.92 and 0.90, respectively, which were significantly better than those achieved by radiologists (resident doctor: 0.66, 0.55, 0.79, 0.74, 0.62 and 0.69; attending doctor: 0.68, 0.59, 0.78, 0.70, 0.69 and 0.73; professor: 0.69, 0.63, 0.72, 0.51, 0.80 and 0.76). Model interpretation with Grad-CAMs revealed that the activated areas mainly focused on cystic necrosis and margins. The US-CNN model predicts GIST malignant potential well, which can assist in clinical treatment decision-making.

Using bedside whiteboard to improve patient-centered care for hospitalized patients in a resource-limited setting: A quality improvement report

Abstract Background In our assessment, hospitalized patients have difficulty of identifying their assigned health-care providers because of frequent rotation of physicians, residents, and interns in a teaching hospital like Tikur Anbessa Specialized Hospital. In addition, patients are not routinely communicated about their daily plan of care and estimated date of discharge. As a solution to these problems, bedside whiteboards could be used with an ultimate goal of improving patient-centered care. Methods Pre-intervention survey was conducted on patients from selected medical wards of the hospital over 3 months (N = 52). Whiteboards were then used on newly admitted patients in the same medical wards, and a post-intervention survey was conducted on these patients over 3 months (N = 47). Responses to the Stanford’s Quality Improvement Team patients’ knowledge assessment and Press Ganey patient satisfaction questionnaires were used as the outcome measures. A pre-/post-intervention study design was used, and all statistical analyses were performed using SPSS Statistical Software 25. Results Most patients were literate, and use of whiteboards significantly increased the proportion of patients who knew their attending doctors and nurses (P &amp;lt; 0.001), daily plans of investigations and treatment options (P &amp;lt; 0.001), and their estimated discharge date (P &amp;lt; 0.001). Patients were also more satisfied with their treatment in all satisfaction assessment questionnaires with significantly higher mean Likert scores in the intervention group compared to the non-intervention group (P ≤ 0.001). Conclusion The findings of this quality improvement project show that bedside whiteboards are effective tools to improve patients’ awareness of their care team, daily care plans, and estimated date of discharge and increase their satisfaction.

Age and gender patterns in health service utilisation: Age-Period-Cohort modelling of linked health service usage records

BackgroundHealth service utilisation changes across the life-course and may be influenced by contextual factors at different times. There is some evidence that men engage less with preventive health services, including attending doctors’ clinics, however the extent to which this varies temporally and across different age groups is unclear. This study aimed to describe age or cohort effects on engagement with GPs among employed mothers and fathers in Australia, and differences in these trends between men and women.MethodsWe linked data from the ‘Growing up in Australia: The Longitudinal Study of Australian Children’ with administrative health service records from Medicare. We used a small-domain estimation Age-Period-Cohort method to describe patterns in health service use among working-age male and female parents in Australia while adjusting for employment status and controlling for time-invariant factors. Our small-domain method assumes a smooth response surface of Age, Period and Cohort.ResultsMale parents have lower health service engagement than women of the same age at the same time period. Men’s pattern of health service use across time is likely explained entirely by ageing. That is, we find that patterns in health service utilisation among men are largely driven by age effects, with no evidence of periods or cohort effects in health service engagement for men between 2002 and 2016.ConclusionsDifferences in health service utilisation between male and female parents at all age-period-cohort combinations highlight a need for more research to examine the extent to which this level of health service use among Australian men meets men’s health needs, as well as barriers and enablers of health service engagement for men. Absence of evidence for period effects suggests that there is little shift in gendered patterns of health service utilisation during the observed period.

Evaluation of the Educational Value of Low-Cost Training Model for Emergency Cricothyrotomy.

In this study, we aimed to evaluate the educational value and students' satisfaction with the hand-made low-cost cricothyrotomy simulation model. A low-cost and hand-made model and a high-fidelity model were used to assess the students. The students' knowledge and satisfaction were evaluated using a 10-item checklist and a satisfaction questionnaire, respectively. Medical interns in the present study participated in a two-hour briefing and debriefing session held in the Clinical Skills Training Center by an emergency attending doctor. Based on the results of data analysis, no significant differences were found between the two groups in terms of gender, age, the month of internship, and last semester's grade (P = .628, .356, .847, and .421, respectively). We also found no significant differences between our groups in terms of the median score of each item in the assessment checklist (P = .838, .736, .805, .172, .439, .823, .243, .950, .812, and .756, respectively). The study groups had no significant difference in the median total scores of the checklist as well (P = .504). Regarding the students' satisfaction, our results showed that interns evaluated their experience of the model as positive (median scores of 4 and 5 out of 5). They also gave the hand-made model a median score of 7 in comparison with the high-fidelity model and 8 out of 10 for its usability. The study results showed that a low-cost model could be as effective as an expensive high-fidelity model for teaching the necessary knowledge of cricothyrotomy techniques to medical trainees.

Interview: Timothy Clinton

Interview: Timothy Clinton

Pneumatosis intestinalis with free air in the abdominal cavity caused by nintedanib.

This is the first case confirmed the association between PI development and nintedanib by the reproducibility of PI development. In patients taking a combination treatment with corticosteroid and nintedanib, clinicians should be careful regarding the development of PI although the patient improved only after discontinuation of nintedanib treatment.

Medical Jousting: Going for the kill!

Medical Jousting: Going for the kill!

PS-C36-1: CLINICAL INERTIA IN THE HYPERTENSIVE PATIENTS TREATED BY HYPERTENSION SPECIALISTS

Background: Only about 30% of hypertensive patients in Japan is reported to have properly controlled blood pressure (BP). The clinical inertia (Inertia) seems to be one of the factors behind the inadequate BP control. Purpose: We investigate the achievement rate of the BP lowering target in 2020 for outpatients in the division of hypertension, and confirm the prevalence of inertia as a possible factor for the failure to achieve their BP targets. Methods: Subjects were 517 outpatients (mean age 68.6 ± 13.2 years, 54% of women) who regularly visited the division of hypertension from March to September 2020 and took antihypertensive drugs prescribed by hypertension specialists. We calculated the average of office BP (OBP) in the two visits during the same period, and the average of home BP (HBP) for the three days before each two visits. In addition, we investigated the antihypertensive drugs taken during that period and complications such as diabetes and chronic kidney disease, and collected blood chemistry and urine test data. From these data, as of December 2021, the attending doctors determined whether each patient achieved the BP target according to the Japanese Society of Hypertension 2019 (JSH2019), and chose one reason from the options in the patients who failed to achieve target BP. Results: The target for OBP in all subjects was &lt; 130/80 (48%), &lt; 140/90 (140 group, 15%), &lt; 140/90 mandatory, and &lt; 130/80 if tolerable (37%). The percentage of those who achieved each OBP target was 48% (OBP achievers, N = 246). As a result of investigating the reason for the OBP non-achievement (N = 273) by the attending doctors, 20% answered “judged to be the optimal BP level (Optimal)”, 20% answered “due to the white coat effect ”, and then “Inertia” was the next highest at 14%. Regarding the achievement rate of HBP for OBP achievers, 36% did not measure HBP, 38% achieved &lt; 135/85 or &lt; 130/80 (HBP 135), and 26% (N = 65) did not achieve it. Among the HBP 135 other than the 140 group (N = 66), the unachieved rate of &lt; 125/75 was 58% (N = 38). The reasons for non-achievement of HBP (N = 103 in total), “Optimal” was the most common answer (42%), followed by “Inertia” (27%). Conclusion: Inertia existed as a factor explaining insufficient BP control even in the patients treated by hypertension specialists. From now on, in order to increase the rate of achieving the BP target, we should make an effort to overcome inertia.

Integrating Mental Health Screening into Antenatal Care

Pregnancy demands comprehensive care, focusing on mental well-being and physical health. Depression and anxiety are prevalent during pregnancy, potentially impacting the mother and foetus. Early intervention is crucial, as severe prenatal depression can adversely affect foetal neurocognitive development. Additionally, a significant proportion of postpartum depression cases originate in the antenatal period, with stressors intensifying throughout pregnancy. Routine screening offers a sophisticated safety net for both mother and foetus. However, effective implementation requires collaborative efforts among healthcare stakeholders. The Obstetrics and Gynaecological Distress Measurement Scale (OGDMS) is a promising tool for comprehensive early screening for the Indian population. It evaluates distress across psychological, physiological, and environmental domains. The OGDMS facilitates dialogue between practitioners and patients, driving modifications in treatment protocols. Involving attending doctors and caregivers fosters a more engaged and effective therapeutic journey. While promising, further studies are essential to solidify the OGDMS’s impact. Future advancements in screening methods promise even more targeted and timely results, ensuring optimal care for expectant mothers.

Nurse-led multi-component behavioural activation programme to improve health outcomes in patients with atrial fibrillation: a mixed-methods study and feasibility analysis.

Patients with atrial fibrillation (AF) play passive roles in disease management. This study aimed to examine the feasibility and preliminary effects of an empowerment-based care model, titled 'the nurse-led multi-component behavioural activation (N-MBA) programme', on health-related quality of life, AF knowledge, psychological outcomes, medication adherence, and treatment decision-making in patients with AF. This mixed-methods study comprised a pilot randomized controlled trial and a qualitative study. Patients with AF who had a moderate-to-high risk of stroke but were not prescribed oral anticoagulants were recruited. Forty participants were recruited and randomized in a 1:1 ratio to receive either the N-MBA programme or standard care. The 13-week programme comprised care components that prepared patients for shared decision-making, an empowerment-based educational module on AF self-care, and continuous support through telephone calls. The programme was feasible, and the overall attendance rate was 82.5%. The participants gave excellent ratings in the satisfaction survey. The N-MBA group showed greater improvements in health-related quality of life (HRQoL) and AF knowledge than the standard care group at the immediate post intervention and 6-month follow-up time points. No significant between-group changes in medication adherence, anxiety, and depression were detected. Participants in the N-MBA group actively raised concerns about AF and its treatment with their attending doctors. The qualitative data were consistent with the quantitative data, indicating that the programme built a comprehensive knowledge base of AF and self-care behaviours. The N-MBA programme is feasible and acceptable to patients with AF. It improved patients' AF knowledge, treatment-related decision-making, and HRQoL. ClinicalTrials.gov NCT03924739.

An exploratory project to develop an effective educational system to teach mastery of assistant laparoscopic gynecologic surgery skills.

In laparoscopic surgery, the cooperation of the first assistant surgeon is essential for the creation and maintenance of an appropriate and safe surgical field. The aim of this study is a validation of the impact of our educational system for first assistant laparoscopic surgeon residency in a single university-affiliated teaching hospital. Five residents were recruited for this study. We created a slide-set as an educational tool using Microsoft's PowerPoint that was in line with our surgical procedure to master the assistant skills of laparoscopic gynecologic surgery. Immediately after surgery, the attending doctor and residents reviewed videos of the first assistant's operative procedures. This study evaluated the effect of our educational checklist against the transition of evaluation scores before and after training in two groups of residents, those with minimal experience (Group 1) or with high experience (Group 2). Before receiving our training tool, there was an expected significant difference in the review scores of the two experience groups (29.1 ± 3.1 vs 48.7 ± 3.9, P=.01). After our modified training, the inexperienced residents of Group 1 received review scores comparable to or higher than those of the experienced residents of Group 2 (81.9 ± 5.9 vs 70.4 ± 7.5, P=.23). As they assisted with more surgeries, the concordance between the resident's self-assessment and the attending doctor's assessment scores significantly increased (P < .001). Our educational system for first assistant surgeons is effective for new residents, as it proved capable of imparting them with many of the same skill sets as much more experienced attending doctors.

Endoscopic photoacoustic with deep learning for evaluating early detection of the gastrointestinal

Colorectal cancer is increasing rapidly every year. At present, traditional medical endoscopic probes are large in size, slow in imaging, and insufficient resolution. Here, we propose a system combining a customized probe-based photoacoustic system with deep learning to improve photoacoustic hardware to achieve a smaller size of the probe and a higher imaging resolution. On the other hand, deep learning was imbedded into the proposed photoacoustic system to achieve highly accurate results of the classifications for potentially helping physicians to identify lesions as the second opinion. In this study, a customized probe with a diameter of 9 mm was used to replace the original probe with a diameter of 1 cm in the hospital. The laser provides 15 W average power with an approximate 300 μJ pulse energy. The laser beam is formed by the lens after focusing, and the fiber couple can convert the light into parallel light. After the probe rotates 360°, a slip ring will trigger the ultrasound to receive particle displacements caused by the thermal propagations. A GRIN lens with a diameter of 1 mm was used to focus the scattered light. The proposed system can generate an 800 μm resolution. The collected images are classified by deep learning algorithms, including AlexNet, GoogLeNet, and ResNet, to differentiate polyps from tumors. After comparing these four image classification methods, ResNet_18 is finally used for image classification, which helps the attending doctor reduce fatigue and quickly identify a disease.