LSLisha Sun

Lisha Sun

*This is an algorithmically generated author record based on publicly available information.

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Lisha Sun has a total of 86 co-authors. Here are the top co-authors by number of shared publications:

  • Tie-Jun Li

    Department of Oral Pathology Peking University School and Hospital of Stomatology Beijing China

    14 shared publications
  • Jiang Xue

    8 shared publications
  • Zhi-Peng Sun

    Research Unit of Precision Pathologic Diagnosis in Tumors of the Oral and Maxillofacial Regions, Chinese Academy of Medical Sciences (2019RU034), Beijing, China

    8 shared publications
  • Aobo Zhang

    Department of Oral Pathology Peking University School and Hospital of Stomatology & National Center of Stomatology & National Clinical Research Center for Oral Diseases & National Engineering Research Center of Oral Biomaterials and Digital Medical Devices Beijing China

    5 shared publications
  • Wenyi Zhang

    4 shared publications
  • Xuefen Li

    Central Laboratory Peking University School and Hospital of Stomatology & National Center of Stomatology & National Clinical Research Center for Oral Diseases & National Engineering Research Center of Oral Biomaterials and Digital Medical Devices Beijing China

    4 shared publications

Latest Publications

  • Deep Learning on Histology Images for Differentiating Fibro-osseous Lesions of the Jaw

    by A.B. Zhang, P.Y. Li, J. Xue, J.Y. Zhang, Z. You, S.H. Ge, Z.X. Xu, Z.P. Sun, D.X. Chang, L.S. Sun, et al.

    Fibrous dysplasia (FD), cemento-ossifying fibroma (COF), and cemento-osseous dysplasia (COD) are fibro-osseous lesions of the jaw that share histologic features, making their differentiation in routine pathology practice challenging. Distinguishing these conditions is crucial for clinical management due to their varying prognoses. To address these limitations, we developed and validated a deep learning model using 1,218 hematoxylin and eosin whole slide images from 338 cases (FD, n = 130; COF, n = 146; COD, n = 62) across 3 institutions. After digital scanning and image preprocessing, we compared 4 training strategies using a ResNet-50 backbone with loss functions and multiple-instance learning: weakly and fully supervised models on single or multiple slides. In the test set, the weakly supervised model on multislide achieved the best performance (area under the curve, 0.86; accuracy, 0.71) as compared with other models, exceeding the diagnostic accuracy of experienced oral pathologists. Visualization with probability heat maps revealed that the model identified key histomorphologic patterns critical for distinguishing FD, COF, and COD. These results demonstrate that a multislide, weakly supervised deep learning approach can provide a classification aid for fibro-osseous lesions with higher discriminative performance than oral pathologists when only histologic slides are considered, highlighting its potential as a supportive tool alongside clinical, radiologic, and molecular data. However, we acknowledge that the test cohort is limited in sample size and geographic diversity; thus, while the results are promising, further expansion and more diverse cohorts would be needed to fully substantiate the generalizability of the model.


  • Open Access

    Bilateral Lesions Are Linked to Postoperative Regrowth in Craniofacial Fibrous Dysplasia: Alkaline Phosphatase as a Marker of Clinical Phenotype

    by Jiang Xue, Longping Liu, Jianyun Zhang, Yue Lou, Lisha Sun, Tiejun Li

    Objectives: This study aimed to evaluate phenotypic associations between preoperative alkaline phosphatase (ALP) levels and clinical characteristics, and explore clinical factors associated for postoperative regrowth in craniofacial fibrous dysplasia. Methods: In this retrospective cohort (2003–2024), 71 surgically treated fibrous dysplasia patients were analyzed. Relationships between preoperative ALP (using age-stratified reference ranges) and key phenotypes (age at surgery, onset age, laterality, lesion type) were assessed via nonparametric tests. Associations with postoperative regrowth were assessed using Mann–Whitney U or Kruskal–Wallis tests for non-normally distributed continuous variables and χ2 or Fisher’s exact tests for categorical variables. Results: Preoperative ALP levels significantly correlated with younger surgical age (16–19 vs. ≥19 years: 244.0 vs. 107.0 U/L, p < 0.001), earlier onset (0–16 vs. >16 years: 114.0 vs. 83.0 U/L, p = 0.030), bilateral lesions (176.0 vs. 106.2 U/L, p = 0.006), and polyostotic subtype (polyostotic fibrous dysplasia vs. monostotic fibrous dysplasia: 162.0 vs. 87.5 U/L, p < 0.001). However, neither ALP levels (p = 0.061) nor abnormal ALP rates (p = 0.090) predicted regrowth. Crucially, bilateral lesions were significantly associated with regrowth (83.3% (5/6) vs. 21.5% (14/65); p = 0.005). The overall regrowth rate was 8.5% (6/71). Conclusions: Bilateral lesions demonstrate significant association with postoperative regrowth risk, potentially guiding surveillance intensity. ALP correlates with phenotypic burden but shows limited prognostic utility. These findings, interpreted considering retrospective constraints, warrant validation in larger cohorts.

  • Open Access

    Performance of Node-RADS for Standardized Diagnosis of Cervical Lymph Nodes in Oral Squamous Cell Carcinoma

    by Jitao Zhu, Wenyi Zhang, Junru Zhao, Bo Feng, Lisha Sun, Zhipeng Sun

    Objectives Current imaging techniques for diagnosing cervical lymph node metastasis (LNM) in oral squamous cell carcinoma (OSCC) lack consensus. The Node-RADS (Node Reporting and Data System) offers a criterion to evaluate LNM based on size and configuration. This study aimed to explore the correlation between Node-RADS scores and LNM rates in OSCC, and assess diagnostic performance. Materials and Methods A retrospective analysis was conducted on 200 patients (average age, 57.8 years, 73.5% were male) diagnosed with OSCC at the Peking University School of Stomatology, who underwent neck dissection, preoperative contrast enhanced CT scans, and had definitive postoperative pathological results. The correlation between Node-RADS scores and LNM rates was examined, various cutoff points (> 1,>2,>3,>4) was applied to evaluate sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV). The ROC and AUC were used to assess diagnostic efficacy. Additionally, the optimal short-axis diameter length and each configuration criterion were determined, while Kappa statistics evaluated inter-rater reliability. Results There was a notable linear correlation between Node-RADS scores and LNM rates. With increasing Node-RADS cutoff values, specificity and PPV increased from 5.6% to 97.2% and 65.3% to 97.5%, sensitivity and NPV decreased from 100% to 60.2% and 100% to 57.9%. The most effective Node-RADS cutoff values for individual patient, lymph node laterality and lymph node level dimension were respectively identified as > 3/>3/>2, with corresponding AUCs of 0.86/0.86/0.82. The Kappa consistency test yielded a value of 0.814. Conclusions This study establishes a groundwork for implementing the Node-RADS in assessing LNM in OSCC, also suggests it has moderate to high accuracy and robust diagnostic performance. Clinical Relevance: The results of this study suggest that Node Reporting and Data System is a reliable tool for guiding clinical decisions in the management of oral squamous cell carcinoma.

  • EGFR T790M Mutations in Conventional and Langerhans Cell‐Rich Variants of CEOT : Unveiling Molecular Differences and Clinical Implications

    by Wenyi Zhang, Jiang Xue, Lisha Sun, Zhipeng Sun

    Objectives: We aimed to investigate the clinicopathologic and molecular characteristics of conventional and Langerhans cell‐rich (LC‐rich) variants of Calcifying Epithelial Odontogenic Tumor (CEOT). Methods: Fifteen patients, eight with conventional and seven with LC‐rich variants, were analyzed using clinical data, radiographic features, and histopathologic assessments. A 560‐gene cancer panel, utilizing Agilent probe hybrid capture technology, was employed for next‐generation sequencing (NGS) to detect genetic mutations. The EGFR c.2369C>T (p.T790M) mutation was identified and further validated using droplet digital PCR (ddPCR). Results: Our findings revealed that conventional CEOT primarily occurred in the mandible, with 75% displaying the EGFR c.2369C>T (p.T790M) mutation. Conversely, the LC‐rich variant was more prevalent in the maxilla, with 28.57% exhibiting the same mutation. Radiographic patterns showed distinct differences, with the LC‐rich variant demonstrating more aggressive clinical behavior. Conclusions: The identification of EGFR c.2369C>T (p.T790M) mutations in both variants underscores its significant role in CEOT pathogenesis and suggests potential implications for targeted therapies. This comprehensive analysis enhances our understanding of the molecular landscape of CEOT variants and highlights the necessity for tailored diagnostic and treatment strategies.


  • Genotype–Phenotype Correlation in Fibrous Dysplasia/ McCune ‐Albright Syndrome Patients With Craniofacial Lesions

    by Jiang Xue, Xuefen Li, Yanrui Feng, Lisha Sun, Tiejun Li

    Objectives: We aimed to investigate genotype–phenotype correlations, variant prevalence, and prognostic factors in a Chinese cohort with craniofacial FD/MAS. Methods: A retrospective study of 93 histologically confirmed FD/MAS cases (2003–2024) analyzed GNAS mutations via direct sequencing. Clinical data, including disease onset, lesion activity (active vs. stable), and extraskeletal manifestations, were correlated with genotypic findings. Results: R201H was the predominant variant (58.1% vs. 41.9% R201C), with no significant differences in demographics, symptoms, or prognosis between genotypes. Active lesions (34.6%) exhibited earlier disease onset (8.0 vs. 12.0 years, p = 0.026), higher rates of bilateral involvement (51.9% vs. 27.5%, p = 0.033), pain (37.0% vs. 9.8%, p = 0.004), and nasal obstruction (25.9% vs. 2.0%, p = 0.003). MAS was strongly associated with active lesions (25.9% vs. 3.9%, p = 0.026), while monostotic FD predominated in stable lesions. All three cases of osteosarcoma harbored R201C; though statistical significance was not reached. Conclusions: This study showed no genotype–phenotype correlations in FD/MAS but highlights R201H as the major variant in this population. Active lesions correlate with early onset, bilateral craniofacial involvement, and MAS, warranting close monitoring. R201C's link to malignancy, though inconclusive, suggests possible variant‐specific differences in malignant transformation.

  • Clinical Characteristics and Prognostic Factors of 199 Salivary Acinic Cell Carcinoma

    by Lijing Zhu, Lisha Sun, Tie‐Jun Li

    Objects: This retrospective study aimed to investigate the clinical characteristics and prognostic features of acinic cell carcinoma (AciCC) of the salivary gland. Methods: One Hundred Ninety‐nine cases of AciCC were retrospectively reviewed. Clinical characteristics and outcomes were systematically collected. Recurrence‐free survival, metastases‐free survival, and overall survival of patients with different clinical variables were compared using the Kaplan–Meier method and Cox univariate regression. Results: A total of 199 patients were included (70 men, 129 women; mean age, 44.57 years). The parotid gland was most frequently affected (75.0%), and the majority of patients were asymptomatic (71.36%). Most patients underwent surgical treatment. Follow‐up data were available for 155 patients, with a mean duration of 78.42 months. Local recurrence occurred in 11.6% of these patients, regional metastases in 6.5%, and distant metastases in 5.2%. Of the 10 patients with metastases, 4 died of AciCC. Poorer survival was significantly associated with age > 40 years, perineural invasion, neck dissection, and advanced stage (III/IV). Recurrence and metastases were strongly associated with AciCC‐related mortality. Conclusions: Regular follow‐up and careful lymph node assessment are critical for patients over 40 years, those with perineural invasion, advanced‐stage disease, recurrence, or metastases, as these factors are linked to poorer outcomes in AciCCs.


  • Mutational Landscape of Ameloblastoma: Analysis of BRAF and Other Key Mutations in Chinese Patients

    by Zhu You, Zihan Sima, Xiaowen Guo, Xuefen Li, Yanrui Feng, Ning Du, Heyu Zhang, Lisha Sun

    Objectives: To investigate the prevalence of BRAF, SMO, KRAS, HRAS, NRAS, FGRF2, and CTNNB1 gene mutations in Chinese ameloblastoma (AM) patients and explore their associations with clinical characteristics.Materials and Methods: DNA was extracted from 89 formalin‐fixed paraffin‐embedded AM samples (9 unicystic and 80 conventional). PCR and Sanger sequencing were used to detect mutations, followed by statistical analysis to assess correlations between mutations and clinical variables.Results: BRAF V600E mutations were significantly prevalent, occurring in 92% (59/64) of mandibular AMs compared to 40% (10/25) in maxillary AMs. SMO mutations were found in 20% (5/25) of maxillary and 3.1% (2/64) of mandibular AMs. FGFR2 mutations were detected in six maxillary and two mandibular AMs, while RAS mutations were present in four maxillary and one mandibular AM. No detectable HRAS, NRAS(G12), or CTNNB1 mutations were observed. BRAF mutations showed mutual exclusivity with SMO and FGFR2 mutations.Conclusion: The high prevalence of BRAF V600E mutations, particularly in mandibular AMs, suggests its potential as a diagnostic and therapeutic target. Distinct mutation profiles between maxillary and mandibular AMs indicate molecular diversity. In BRAF‐negative cases, alternative oncogenic pathways involving SMO, FGFR2, and RAS may be actionable targets, underscoring the need for personalized treatment approaches.


  • Identification of Novel and Rare GNAS Mutations in Craniofacial Fibrous Dysplasia

    by Jiang Xue, Jianyun Zhang, Ming Ma, Xuefen Li, Lisha Sun, Ruirui Shi, Tiejun Li

    Background: Fibrous dysplasia (FD), caused by activating mutations of GNAS, is a skeletal disorder with considerable clinicopathological heterogeneity. Although prevalent mutations such as R201C and R201H dominate in FD, a limited number of rare mutations, including R201S, R201G, and Q227L, have been documented. The scarcity of information concerning these uncommon mutations motivates our investigation, seeking to enhance comprehension of this less‐explored subgroup within FD.Methods: This study introduces three cases of craniofacial FD exhibiting rare GNAS mutations. Employing DNA sequencing on fresh frozen lesion tissues, we conducted a thorough analysis of clinical, radiological, and pathological features, delving into genotypic and phenotypic correlations. A comparative assessment with our previous series was also conducted.Results: In the subset of patients subjected to DNA sequencing, a novel GNAS missense mutation (Q227E) was identified in one case, while a rare GNAS mutation (R201S) in exon 8 was found in the other two patients. Although no apparent phenotypic distinctions were observed among those with GNAS hotspot mutations (R201C, R201H), a more severe phenotype was discerned in the case featuring the novel GNAS mutation Q227E.Conclusions: This study marks the first report of the Q227E mutation in the GNAS gene associated with bone disease, enriching our understanding of FD's genetic basis and shedding light on the clinicopathological heterogeneity of craniofacial FD.