Automatic segmentation of ameloblastoma on ct images using deep learning with limited data.

BACKGROUND: Ameloblastoma, a common benign tumor found in the jaw bone, necessitates accurate localization and segmentation for effective diagnosis and treatment. However, the traditional manual segmentation method is plagued with inefficiencies and drawbacks. Hence, the implementation of an AI-based automatic segmentation approach is crucial to enhance clinical diagnosis and treatment procedures. METHODS: We collected CT images from 79 patients diagnosed with ameloblastoma and employed a deep learning neural network model for training and testing purposes. Specifically, we utilized the Mask R-CNN neural network structure and implemented image preprocessing and enhancement techniques. During the testing phase, cross-validation methods were employed for evaluation, and the experimental results were verified using an external validation set. Finally, we obtained an additional dataset comprising 200 CT images of ameloblastoma from a different dental center to evaluate the model's generalization performance. RESULTS: During extensive testing and evaluation, our model successfully demonstrated the capability to automatically segment ameloblastoma. The DICE index achieved an impressive value of 0.874. Moreover, when the IoU threshold ranged from 0.5 to 0.95, the model's AP was 0.741. For a specific IoU threshold of 0.5, the model achieved an AP of 0.914, and for another IoU threshold of 0.75, the AP was 0.826. Our validation using external data confirms the model's strong generalization performance. CONCLUSION: In this study, we successfully applied a neural network model based on deep learning that effectively performs automatic segmentation of ameloblastoma. The proposed method offers notable advantages in terms of efficiency, accuracy, and speed, rendering it a promising tool for clinical diagnosis and treatment.

Dentigerous cysts suspected the other odontogenic lesions on panoramic radiography and CT.

Dentigerous cysts are known as the second most common type of cyst in the jaws. The cyst is one of the lesions occurred frequently in the posterior body of the mandible and is often related to the unerupted third molar and forms around the crown of the unerupted tooth attaching at the cementoenamel junction. Such characteristic appearances are the diagnostic points differentiating from ameloblastoma or odontogenic keratocyst. However, it would be hard for us to diagnose it as a dentigerous cyst if the lesion does not show its typical appearance. We experienced two cases of dentigerous cysts which did not form around the crown of the unerupted tooth on radiologically. Both cysts were relatively large and resorbed adjacent teeth roots. Therefore, an ameloblastoma or an odontogenic keratocyst was suspected rather than a dentigerous cyst as the imaging diagnosis. The biopsy revealed that the lesion was a "dentigerous cyst" in one of the cases and "developmental cyst with inflammation" in another case. After the excision, the histopathological diagnosis was a dentigerous cyst with inflammation in both cases. This report shows the two cases of dentigerous cysts focusing on panoramic radiography and CT images. Also, we discuss the differential diagnosis by reconsidering those diagnostic points.

Clinicopathologic Profile and Surgical Modalities in Mandibular Ameloblastoma: A Descriptive Study.

BACKGROUND: Ameloblastoma is a benign neoplasm composed of epithelial tissue with invasive and infiltrative behavior at the local level and a high recurrence rate, with various histopathologic patterns and clinical forms. Approximately 85% of conventional ameloblastomas occur in the mandible, most often in the body, angle, and ascending ramus area. The treatment modalities include both conservative and radical treatments. Postoperative follow-up is most important in the treatment of ameloblastoma. AIMS AND OBJECTIVES: To describe the clinicopathologic profile of mandibular ameloblastoma in patients undergoing different surgical modalities. The primary objective was to describe the clinicopathologic profile and surgical management of mandibular ameloblastoma in patients aged ≥18 years, who had reported to a tertiary dental care center for follow-up during the study period. The secondary objective was to describe the distribution of comorbidities associated with different surgical modalities and reconstructive methods. SUBJECTS AND METHODS: A total of 34 patients with mandibular ameloblastoma who underwent various surgical modalities between 2011 and 2021 were studied. Information was collected using a predesigned proforma and statistically analyzed. RESULTS: Thirty-four review cases of ameloblastoma were included in the study. The patients were analyzed concerning age, sex, site, size, clinical presentation, radiographic pattern, histopathologic subtype, type of surgery, and associated comorbidities. Most cases of mandibular ameloblastoma involve the age of 16 to 55 years. The mean age of occurrence was found to be 35.5±13.2. A female preponderance, a tumor size range of 2 to 4 cm, a multicystic variant, involvement of the mandibular body in the premolar-molar area, root resorption, cortical perforation, and a follicular type of histopathologic pattern were the common presentations. Isolated anterior tumors restricted to the incisor/canine region were not found. The common surgical modalities undertaken were conservative methods such as enucleation, and chemical cauterization, and radical methods such as marginal mandibulectomy and segmental resection. Reconstruction using a titanium plate or free fibular graft was performed in the indicated cases. The common comorbidities included difficulty in chewing and loss of facial contour. Recurrence after surgical treatment was rare. Only 9% of cases developed a recurrence within 5 years. No recurrence was noted in cases treated with radical treatment, whereas 50% of cases treated with conservative methods showed recurrence. CONCLUSION: The age of occurrence, site, and size of the tumor, cortical perforation, root resorption, histopathologic type, and radiographic patterns are widely considered factors in devising a treatment plan for mandibular ameloblastoma. However, there may be rare instances where these tumors behave differently regardless of their innocuous clinicopathologic presentation. Surgical procedures such as segmental resection and marginal mandibulectomy were found to be promising for the eradication of the tumor, and prevention of recurrences and metastasis. However, conservative measures such as enucleation and chemical cauterization were fraught with an increased risk of tumor recurrence and metastasis. Future studies with a larger sample size should focus on the clinicopathologic characteristics of ameloblastoma to elucidate its varied behavior and develop newer and advanced treatment modalities that would provide better surgical and postsurgical outcomes in affected patients.

Peripheral ameloblastoma underlying squamous cell papilloma after a third molar extraction.

Peripheral ameloblastoma (PA) is a rare variant of ameloblastoma that presents as a slow-growing, painless mass in the gingival tissues or alveolar mucosa. It shares histologic features with conventional ameloblastoma but is less invasive and aggressive. This case report describes a 51-year-old female with a PA that simultaneously or subsequently developed underlying squamous cell papilloma after mandibular third molar extraction. Clinical examination revealed a pedunculated gingival lesion mimicking squamous cell papilloma. Histopathologic examination confirmed PA underlying squamous cell papilloma after an excisional biopsy. Imaging revealed mild bone resorption, leading to a further soft tissue excision and minimal osteoectomy to rule out intraosseous involvement. The patient remained asymptomatic without signs of recurrence in the 1-year follow-up. PA diagnosis can be challenging due to its clinical resemblance to other gingival lesions and histopathologic features. Treatment typically involves surgical excision, with long-term follow-up recommended due to possible recurrence and malignant transformation.

[Cone-beam CT imaging features of common cystic lesions associated with the impacted mandibular third molar].

Objective: To analyze the imaging features of cone-beam CT (CBCT) of ameloblastoma (AB), odontogenic keratocyst (OKC) and dentigerous cysts (DC) associated with the mandibular impacted third molars,so as to provide useful information for differential diagnosis of these lesions. Methods: The patients who were with complete clinical data, pathological diagnosis and CBCT images from AB, OKC and DC around the mandibular impacted third molars were collected in Peking University Hospital of Stomatology from August 2016 to December 2021. A total of 109 patients (14 were diagnosed as AB, 23 were diagnosed as OKC and the others were diagnosed as dental cysts) were collected, including 73 males and 36 females. The age ranged from 11 to 70. The analyzed imaging features included location and internal density of the lesions, bone expansion, root resorption of adjacent teeth and types of the impacted teeth. The Chi square test was used to compare the gender of different lesions, and the Fisher's exact test was used to compare imaging features of lesions. When P<0.05, there was a significant difference among the three. Logistic regression analysis was performed to determine the imaging features that significantly contribute to correct imaging diagnosis. Corresponding P-values were calculated for all factors from multivariate models. Results: In the 23 cases of OKC, no special location was observed for the center of the lesion, heterogeneous high-density were seen in 21.7% of the cases, 56.5% of the cases had no significant bone expansion and the impacted teeth were not specially oriented. Among the 14 AB, 7 cases (7/14) were mainly located in the ramus of the mandible, and all cases (14/14) had buccal/lingual expansion of the jaw, 8 cases (8/14) presented root resorption of the adjacent teeth, and mesial impacted mandibular third molar were seen in 6 cases (6/14). Among the 72 DC, 88.9% (64/72) of the cases were mainly limited to the crown of the impacted third molar, 72.2% (52/72) of the cases had no obvious bone expansion, inverted impacted teeth were shown in 56.9% (41/72) of the cases. There was a significant difference among the three groups (χ2=7.30, P=0.026) in gender. AB and odontogenic cyst were more common in men than in women, while the incidence of OKC was roughly equal between men and women.There were significant differences in the location (P<0.001), internal density (P=0.001) of the lesions, bone expansion (P<0.001) and types of the impacted teeth (P<0.001), while no statistical difference was found for root resorption of adjacent teeth (P=0.153). Logistics regression analysis showed that the location of the lesion, internal density, bone expansion, root resorption of adjacent teeth and the types of impacted teeth had significant effects on the accurate diagnosis of the three kinds of lesions. Conclusions: Location, internal density, bone expansion and types of the impacted teeth played an important role in the correct imaging diagnosis. Further analysis indicates that when the classification of impacted teeth and the location of lesions are considered synchronously, DC can be differentiated from AB and OKC.

Deep learning applied to the histopathological diagnosis of ameloblastomas and ameloblastic carcinomas.

BACKGROUND: Odontogenic tumors (OT) are composed of heterogeneous lesions, which can be benign or malignant, with different behavior and histology. Within this classification, ameloblastoma and ameloblastic carcinoma (AC) represent a diagnostic challenge in daily histopathological practice due to their similar characteristics and the limitations that incisional biopsies represent. From these premises, we wanted to test the usefulness of models based on artificial intelligence (AI) in the field of oral and maxillofacial pathology for differential diagnosis. The main advantages of integrating Machine Learning (ML) with microscopic and radiographic imaging is the ability to significantly reduce intra-and inter observer variability and improve diagnostic objectivity and reproducibility. METHODS: Thirty Digitized slides were collected from different diagnostic centers of oral pathology in Brazil. After performing manual annotation in the region of interest, the images were segmented and fragmented into small patches. In the supervised learning methodology for image classification, three models (ResNet50, DenseNet, and VGG16) were focus of investigation to provide the probability of an image being classified as class0 (i.e., ameloblastoma) or class1 (i.e., Ameloblastic carcinoma). RESULTS: The training and validation metrics did not show convergence, characterizing overfitting. However, the test results were satisfactory, with an average for ResNet50 of 0.75, 0.71, 0.84, 0.65, and 0.77 for accuracy, precision, sensitivity, specificity, and F1-score, respectively. CONCLUSIONS: The models demonstrated a strong potential of learning, but lack of generalization ability. The models learn fast, reaching a training accuracy of 98%. The evaluation process showed instability in validation; however, acceptable performance in the testing process, which may be due to the small data set. This first investigation opens an opportunity for expanding collaboration to incorporate more complementary data; as well as, developing and evaluating new alternative models.

Machine learning-based radiomics for predicting BRAF-V600E mutations in ameloblastoma.

Background: Ameloblastoma is a locally invasive and aggressive epithelial odontogenic neoplasm. The BRAF-V600E gene mutation is a prevalent genetic alteration found in this tumor and is considered to have a crucial role in its pathogenesis. The objective of this study is to develop and validate a radiomics-based machine learning method for the identification of BRAF-V600E gene mutations in ameloblastoma patients. Methods: In this retrospective study, data from 103 patients diagnosed with ameloblastoma who underwent BRAF-V600E mutation testing were collected. Of these patients, 72 were included in the training cohort, while 31 were included in the validation cohort. To address class imbalance, synthetic minority over-sampling technique (SMOTE) is applied in our study. Radiomics features were extracted from preprocessed CT images, and the most relevant features, including both radiomics and clinical data, were selected for analysis. Machine learning methods were utilized to construct models. The performance of these models in distinguishing between patients with and without BRAF-V600E gene mutations was evaluated using the receiver operating characteristic (ROC) curve. Results: When the analysis was based on radiomics signature, Random Forest performed better than the others, with the area under the ROC curve (AUC) of 0.87 (95%CI, 0.68-1.00). The performance of XGBoost model is slightly lower than that of Random Forest, and its AUC is 0.83 (95% CI, 0.60-1.00). The nomogram evident that among younger women, the affected region primarily lies within the mandible, and patients with larger tumor diameters exhibit a heightened risk. Additionally, patients with higher radiomics signature scores are more susceptible to the BRAF-V600E gene mutations. Conclusions: Our study presents a comprehensive radiomics-based machine learning model using five different methods to accurately detect BRAF-V600E gene mutations in patients diagnosed with ameloblastoma. The Random Forest model's high predictive performance, with AUC of 0.87, demonstrates its potential for facilitating a convenient and cost-effective way of identifying patients with the mutation without the need for invasive tumor sampling for molecular testing. This non-invasive approach has the potential to guide preoperative or postoperative drug treatment for affected individuals, thereby improving outcomes.

Unicystic ameloblastoma - A case series.

Ameloblastoma is a benign, locally aggressive neoplasm that constitutes about 1-3% of the tumors of the jaw. Wide surgical excision with adequate safe margin is the most common treatment of choice. The study aimed to manage cases with unicystic ameloblastoma while preserving the continuity of the mandible (without resection). This article presents a series of cases ranging from 18 to 40 years old patients of both sexes with unicystic ameloblastoma, especially in the mandible showing more male predilection than female. All the cases presented in this article were treated by enucleation and curettage. None of the patients presented post-operative paresthesia. None of the cases went in for resection. Post-operative recovery was uneventful in all the patients. All the patients were followed up for a period of 3.5-5 years. None of the cases reported recurrence at the date of publication.

Mandibular reconstruction after excision of recurrent odontogenic keratocyst using a novel mandibular distraction osteogenesis method- a case report.

BACKGROUND: Odontogenic keratocyst is one of the most common benign odontogenic neoplasms with a high recurrence rate. Its resection has the potential to lead to mandibular segmental defects. In this case report, we describe a patient with odontogenic keratocyst who underwent radical resection using a novel distraction osteogenesis (DO) method to reconstruct mandibular segmental defect. CASE PRESENTATION: This case report describes a 19-year-old woman with odontogenic keratocyst of the mandible that recurred after multiple curettages and eventually necessitated radical resection. Mandibular segmental defect after radical resection was reconstructed using a novel DO method that involved directly contacting the segment ends of the defect without the transport disk. However, the distractor broke during the retention period, and a molding titanium plate was used for fixation. This novel distraction method achieved mandibular reconstruction and restored mandibular function and contour.

Dentinogenic ghost cell tumor associated with odontoma: report of a rare case and review of literature.

Dentinogenic ghost cell tumor (DGCT) is an uncommon odontogenic tumor considered to be the solid, tumorous variant of calcifying odontogenic cyst. DGCT is characterized by islands of ameloblastoma-like epithelial cells that resemble the enamel organ, the presence of ghost cells, and dentinoid material. This article reports a rare case of dentinogenic ghost cell tumor associated with an odontoma in an adult patient, with a review of the literature. To the best of the authors' knowledge, there have only been four case reports of DGCT associated with odontoma, all of which occurred in children and adults younger than 30 years old.

[Application of Deep Learning in Differential Diagnosis of Ameloblastoma and Odontogenic Keratocyst Based on Panoramic Radiographs].

Objective To evaluate the accuracy of different convolutional neural networks (CNN),representative deep learning models,in the differential diagnosis of ameloblastoma and odontogenic keratocyst,and subsequently compare the diagnosis results between models and oral radiologists. Methods A total of 1000 digital panoramic radiographs were retrospectively collected from the patients with ameloblastoma (500 radiographs) or odontogenic keratocyst (500 radiographs) in the Department of Oral and Maxillofacial Radiology,Peking University School of Stomatology.Eight CNN including ResNet (18,50,101),VGG (16,19),and EfficientNet (b1,b3,b5) were selected to distinguish ameloblastoma from odontogenic keratocyst.Transfer learning was employed to train 800 panoramic radiographs in the training set through 5-fold cross validation,and 200 panoramic radiographs in the test set were used for differential diagnosis.Chi square test was performed for comparing the performance among different CNN.Furthermore,7 oral radiologists (including 2 seniors and 5 juniors) made a diagnosis on the 200 panoramic radiographs in the test set,and the diagnosis results were compared between CNN and oral radiologists. Results The eight neural network models showed the diagnostic accuracy ranging from 82.50% to 87.50%,of which EfficientNet b1 had the highest accuracy of 87.50%.There was no significant difference in the diagnostic accuracy among the CNN models (P=0.998,P=0.905).The average diagnostic accuracy of oral radiologists was (70.30±5.48)%,and there was no statistical difference in the accuracy between senior and junior oral radiologists (P=0.883).The diagnostic accuracy of CNN models was higher than that of oral radiologists (P<0.001). Conclusion Deep learning CNN can realize accurate differential diagnosis between ameloblastoma and odontogenic keratocyst with panoramic radiographs,with higher diagnostic accuracy than oral radiologists.

A conservative approach for unicystic ameloblastoma: Retrospective clinic-pathologic analysis of 12 cases.

BACKGROUND: Unicystic ameloblastoma is an encapsulated odontogenic neoplasm with a single cyst cavity. The conservative or aggressive surgical approaches used to treat the tumor directly affect recurrence rates. However, there is a lack of a standard protocol that can guide its management. STUDY DESIGN: We retrospectively reviewed the clinicopathological findings and therapeutical procedures of 12 unicystic ameloblastoma cases treated by the same surgeon during the past 20 years. METHODS: All cases of unicystic ameloblastoma diagnosed by biopsy and treated by the same surgeon between 2002 and 2022 were reviewed. Eligibility criteria were patients with completely filled-out charts containing the follow-up period and confirmation of the diagnoses based on the microscopic findings of the whole excised specimens. Data collected were categorized into clinical, radiographic, histological, surgical, and recurrence aspects. RESULTS: There was a female predilection (2:1), and ages ranged between 18 and 61 years (mean: 27.25, ±12.45). Almost all (92%) affected the posterior mandible. Radiographically, the mean length of the lesions was 46.14 mm ± 14.28 mm which 92% were unilocular and 8.3% multilocular. Root resorption (n = 7, 58%), tooth displacement (n = 9, 75%), and cortical perforation (n = 5, 42%) were also observed. The mural histological subtype corresponded to 9 (75%) of the cases. The same conservative protocol was performed in all cases. The follow-up period ranged between 12 and 240 months (~62 ± 65) and recurrence occurred in only one patient (8%). CONCLUSION: Our findings suggest a conservative approach should be the first option for unicystic ameloblastoma treatment, even for those with mural proliferation.

Dentinogenic ghost cell tumor treated with a combination of marsupialization and radical resection: a case report and review of the literature.

BACKGROUND: Dentinogenic ghost cell tumor is a rare benign tumor that accounts for less than 3% of all cases and consists of the stellate reticulum, which is made up of enamel epithelioid and basaloid cells. Although DGCT is a benign tumor, the local infiltration of the odontogenic epithelium or recurrences have been reported, and its detailed pathology and treatments remain unclear. CASE PRESENTATION: This report describes the case of a 60-year-old Japanese male diagnosed with a maxillary dentinogenic ghost cell tumor. Images showed well-circumscribed, multilocular cystic lesions with a calcified substance in the interior. Marsupialization was performed along with biopsy to prevent the expansion of the lesion, and a partial maxillectomy was performed 2 years after the initial examination. Histopathological findings showed ameloblastomatous proliferation containing clusters of ghost cells and dentinoid materials, resulting in the diagnosis of dentinogenic ghost cell tumor. This article also reviews recently reported cases of dentinogenic ghost cell tumor. CONCLUSION: It is important to perform marsupialization, proper resection, and postoperative follow-up because of possible recurrence.

Volumetric change of bony cavity and shrinkage speed after marsupialization for odontogenic keratocyst and unicystic ameloblastoma.

The aim of this study was to evaluate the efficacy of marsupialization treatment for odontogenic keratocyst (OKC) and unicystic ameloblastoma (UA) based on the three-dimensional volumetric change over time, and to determine the difference between OKC and UA in terms of the absolute volume reduction (AVR) and absolute shrinkage speed (ASS), and whether they are correlated with the preoperative volume, time after marsupialization (time between marsupialization and second treatment), and patient age. This was a retrospective cohort study with a sample size of 60 patients: 29 with OKC and 31 with UA. Pre- and post-marsupialization cone beam computed tomography images were analysed using Mimics software. The volume reduction and shrinkage speed were analysed and correlated with the preoperative volume, time after marsupialization, and demographic data. Descriptive univariable and multivariable statistics were computed; significance was set at P ≤ 0.05. The mean percentage volume reduction after marsupialization was 67.6 ± 9.6% for OKC and 63.3 ± 20.1% for UA. There was no significant difference in AVR or ASS between the OKC and UA groups. For OKC and UA, the preoperative volume (both P < 0.001) and time after marsupialization (P = 0.024 and P < 0.001, respectively) were associated with AVR. Moreover, for OKC and UA, the preoperative volume and time after marsupialization were also significantly associated with the ASS (all P < 0.001). For both lesions, patient age was not significantly related to AVR or ASS. Marsupialization appears to be a viable option to decrease the volume of OKC and UA. Age was found not to be associated with the volume reduction of either UA or OKC.

A Rare Case of a Large Ameloblastoma and Cervical Lynphoepitelial Cyst Treated Simultaneously by an Intraoral Approach and Immediate Reconstruction.

Ameloblastoma is a benign locally aggressive odontogenic tumor of epithelial origin with unlimited growth capacity. Cervical lymphoepithelial cyst is a developmental cyst located in the superior-lateral region of the neck with treatment consisting of surgical excision, which presents low recurrence rates. This article reports the case of patient J.G.S.L., 24 years old, female, who reported to the service with a histopathological evaluation compatible with intraosseous mandibular ameloblastoma. On physical examination, a nontender swelling was palpable along the body of left mandible where the ameloblastoma was located, with a fluctuant region detected just below it. Imaging exams confirmed the suspicion of two independent lesions. The treatment plan consisted of two stages: first, mandibular resection of the area affected by the ameloblastoma, followed by reconstruction with a titanium plate simultaneous to excisional biopsy of the as yet undiagnosed lesion, both through intraoral access; second, mandibular reconstruction with iliac crest graft. Histopathological examination of the second lesion indicated a diagnosis of cervical lymphoepithelial cyst, being the first case in the literature with these concomitant lesions. Currently, the patient is in her third postoperative year with no signs of recurrence.

A rare case of peripheral adenoid ameloblastoma with dentinoid.

Adenoid ameloblastoma with dentinoid (AAD) is an extremely rare central neoplasm of the gnathic bones with no reported case of peripheral AAD. Adenoid ameloblastoma with dentinoid was first reported by Slabbert et al. in 1992 under the term "dentinoameloblastoma." However, the Armed Forces Institute of Pathology introduced the name "adenoid ameloblastoma with dentinoid" in 1994. Histologically, AAD shows features of ameloblastoma and adenomatoid odontogenic tumor along with a calcified product resembling dentin. The biological behavior of central AAD is not well established, though, in the literature, several reports consider it to be among the more aggressive odontogenic neoplasms, with a propensity for local invasion and recurrence. The demographic characteristics, clinical features, behavior, and prognosis of the peripheral AAD (PAAD) are unknown. To the best of our knowledge, this is the first reported case of PAAD in a 62-year-old woman.

Clinical outcome following conservative treatment of 58 mandibular ameloblastoma patients: a retrospective study.

The aim of this study was to retrospectively analyze the surgical outcome of the patients with mandibular ameloblastoma (MA) with intact inferior cortex and the lingual cortex of the mandible, treated with conservative management in the form of enucleation with peripheral ostectomy, mechanical curettage, chemical cauterization, dredging, and marsupialization. The patients were included with the diagnosis of MA between the age group of 14 to 60 years with their computed tomography (CT) scans showing intact inferior cortex and the lingual cortex of the mandibular walls. Out of 96 cases of ameloblastoma, 58 (61%) were in mandible and 38 (39%) cases in maxilla. We analyzed these 58 patients retrospectively which were operated conservatively between January 2009 and December 2018. The treatment protocol for all the solid variants and the unicystic variants with intact buccal, inferior, and the lingual cortex was enucleation with peripheral ostectomy followed by mechanical curettage, chemical cauterization, and subsequent dredging. This was performed in 90% (n = 52) cases, whereas the treatment protocol for unicystic variants with perforated buccal cortex and intact inferior and lingual cortex was marsupialization followed by the enucleation with peripheral ostectomy, mechanical curettage, chemical cauterization, and subsequent dredging which was performed in 10% (n = 6) cases. Chi square test was applied to the effectiveness of conservative management as outcome variable. The new bone formation along with bony trabeculae was found in 96.6% (n = 56) of the cases. Recurrence was noted in 3.44% (n = 2) of the cases. A p value was 0.001, thus suggesting statistically highly significant result. On the basis of present study, though we can conclude that the MA with the intact inferior and the lingual cortex can be managed effectively with conservative treatment with long term regular follow up, still we would caution that the reader must use their best clinical judgement based on latest available data. For some patients, returning to clinic frequently and undergoing multiple minor procedures may not be suitable to be considered the best, whereas resection and reconstruction may be deemed more appropriate.

Application of Deep Learning in Differential Diagnosis of Ameloblastoma and Odontogenic Keratocyst Based on Panoramic Radiographs / 中国医学科学院学报

Objective To evaluate the accuracy of different convolutional neural networks (CNN),representative deep learning models,in the differential diagnosis of ameloblastoma and odontogenic keratocyst,and subsequently compare the diagnosis results between models and oral radiologists. Methods A total of 1000 digital panoramic radiographs were retrospectively collected from the patients with ameloblastoma (500 radiographs) or odontogenic keratocyst (500 radiographs) in the Department of Oral and Maxillofacial Radiology,Peking University School of Stomatology.Eight CNN including ResNet (18,50,101),VGG (16,19),and EfficientNet (b1,b3,b5) were selected to distinguish ameloblastoma from odontogenic keratocyst.Transfer learning was employed to train 800 panoramic radiographs in the training set through 5-fold cross validation,and 200 panoramic radiographs in the test set were used for differential diagnosis.Chi square test was performed for comparing the performance among different CNN.Furthermore,7 oral radiologists (including 2 seniors and 5 juniors) made a diagnosis on the 200 panoramic radiographs in the test set,and the diagnosis results were compared between CNN and oral radiologists. Results The eight neural network models showed the diagnostic accuracy ranging from 82.50% to 87.50%,of which EfficientNet b1 had the highest accuracy of 87.50%.There was no significant difference in the diagnostic accuracy among the CNN models (P=0.998,P=0.905).The average diagnostic accuracy of oral radiologists was (70.30±5.48)%,and there was no statistical difference in the accuracy between senior and junior oral radiologists (P=0.883).The diagnostic accuracy of CNN models was higher than that of oral radiologists (P<0.001). Conclusion Deep learning CNN can realize accurate differential diagnosis between ameloblastoma and odontogenic keratocyst with panoramic radiographs,with higher diagnostic accuracy than oral radiologists.