Merve Gursoy
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Merve Gursoy has a total of 76 co-authors. Here are the top co-authors by number of shared publications:
Berna Dirim Mete
Department of Radiology, Faculty of Medicine, Izmir Democracy University, Izmir, Turkey
6 shared publications- Berna Dirim Mete
Faculty of Medicine - Izmir Democracy University, Izmir, Turkey
5 shared publications - Tugrul Bulut
Department of Orthopaedics and Traumatology, Ataturk Training and Research Hospital, Izmir Katip Celebi University, Izmir, Turkey
5 shared publications Tugrul Bulut
Izmir Katip Celebi University, Ataturk Training and Research Hospital, Department of Orthopaedics and Traumatology, Izmir, Turkey
4 shared publicationsFatih Dag
Department of Radiology, Manisa Turgutlu State Hospital, Manisa, Izmir, Turkey
3 shared publications- Atilla Hikmet Cilengir
Faculty of Medicine, Department of Radiology, Izmir Demokrasi University, Izmir, Türkiye
2 shared publications
Latest Publications
- 21 October 2026
- Published by Oxford University Press (OUP) in Journal of Breast Imaging
Benchmark Evaluation of Vision-Enabled Large Language Models in Mammography: A Case–Control Study
Objective: To compare the diagnostic performance of five contemporary vision-enabled large language models (LLMs) in mammography interpretation using a case–control benchmark design. Methods: A total of 200 cases were analyzed using a quota-balanced case–control design (100 malignant, BI-RADS 4–5; 100 controls, BI-RADS 1–2). All models (ChatGPT-5T, ChatGPT-5I, ChatGPT-4o, Gemini 2.5 Pro, and Gemini 2.5 Flash) received standard mammography images. Models were evaluated for lesion detection, laterality, quadrant localization, lesion type, and malignant–benign discrimination. Performance was assessed using accuracy, sensitivity, specificity, balanced accuracy, Cohen’s κ, and area under the receiver operating characteristic curve (AUC). Breast density was recorded for all cases, while histopathological subtype and lesion size were documented for malignant cases. Results: Lesion detection accuracy ranged from 79% to 91%, with high malignant sensitivity (92%–100%) but lower specificity in normal cases (24%–82%). ChatGPT-5T performed best in laterality (83%; κ = 0.70) and quadrant localization (73%; κ = 0.62). Lesion type classification was modest (74%–85%). In malignant–benign discrimination, ChatGPT-5T achieved the highest balanced accuracy (87%), while ChatGPT-4o showed poor specificity (28%). The study cohort was predominantly composed of lower-density breasts, larger lesions, and invasive ductal carcinoma as the most common subtype. Conclusion: Vision-enabled LLMs demonstrated high sensitivity for breast lesion detection and classification tasks. More deliberative models, particularly ChatGPT-5T, consistently outperformed faster counterparts, especially in spatial reasoning tasks. However, reduced specificity in normal cases underscores that current LLM outputs should remain under radiologist supervision as supportive second-reader tools rather than standalone diagnostic systems. Given the quota-balanced design and cohort composition, results should be interpreted as comparative benchmarking rather than real-world diagnostic performance.
- 21 October 2026
- Published by Oxford University Press (OUP) in Journal of Breast Imaging
Benchmark Evaluation of Vision-Enabled Large Language Models in Mammography: A Case–Control Study
Objective: To compare the diagnostic performance of five contemporary vision-enabled large language models (LLMs) in mammography interpretation using a case–control benchmark design. Methods: A total of 200 cases were analyzed using a quota-balanced case–control design (100 malignant, BI-RADS 4–5; 100 controls, BI-RADS 1–2). All models (ChatGPT-5T, ChatGPT-5I, ChatGPT-4o, Gemini 2.5 Pro, and Gemini 2.5 Flash) received standard mammography images. Models were evaluated for lesion detection, laterality, quadrant localization, lesion type, and malignant–benign discrimination. Performance was assessed using accuracy, sensitivity, specificity, balanced accuracy, Cohen’s κ, and area under the receiver operating characteristic curve (AUC). Breast density was recorded for all cases, while histopathological subtype and lesion size were documented for malignant cases. Results: Lesion detection accuracy ranged from 79% to 91%, with high malignant sensitivity (92%–100%) but lower specificity in normal cases (24%–82%). ChatGPT-5T performed best in laterality (83%; κ = 0.70) and quadrant localization (73%; κ = 0.62). Lesion type classification was modest (74%–85%). In malignant–benign discrimination, ChatGPT-5T achieved the highest balanced accuracy (87%), while ChatGPT-4o showed poor specificity (28%). The study cohort was predominantly composed of lower-density breasts, larger lesions, and invasive ductal carcinoma as the most common subtype. Conclusion: Vision-enabled LLMs demonstrated high sensitivity for breast lesion detection and classification tasks. More deliberative models, particularly ChatGPT-5T, consistently outperformed faster counterparts, especially in spatial reasoning tasks. However, reduced specificity in normal cases underscores that current LLM outputs should remain under radiologist supervision as supportive second-reader tools rather than standalone diagnostic systems. Given the quota-balanced design and cohort composition, results should be interpreted as comparative benchmarking rather than real-world diagnostic performance.
- Open Access
Magnetic Levitation of Intact Tumor Biopsies Reveals Multivariate Biophysical Signatures of Breast Cancer Aggressiveness
Diagnostic assessment of breast cancer biopsies remains reliant on resource-intensive histopathology and molecular profiling, which often lack real-time physiological readouts. Magnetic levitation (MagLev) enables label-free density profiling of single cells, yet its application to intact tissue biopsies has been precluded by size-dependent geometric artifacts and the absence of analytical frameworks for biopsy-scale samples. Here, we report the first application of MagLev to intact invasive breast carcinoma biopsies (200-600 μm) for biophysical profiling, generating multivariate biophysical signatures from 203 samples across 17 patients. We developed a physics-based size-correction algorithm (xmc) that isolates biological density from geometric artifact, and demonstrate that tissue viability is predicted not by average levitation height, but by spatial heterogeneity across replicate samples, reflecting the microenvironmental complexity of metabolically active tumors. Multivariate integration using Partial Least Squares (PLS) regression and Factor Analysis of Mixed Data (FAMD) identified nodal status (N) as the strongest biophysical predictor, suggesting that lymphatic dissemination capacity leaves a measurable signature in the primary tumor density profile. Unsupervised patient clustering in PLS-derived latent space recovered three clinically coherent subgroups aligned with molecular subtypes. This 30-minute, low-cost assay provides exploratory biophysical stratification complementary to existing diagnostics, particularly in resource-limited settings.
Open AccessMagnetic Levitation of Intact Tumor Biopsies Reveals Multivariate Biophysical Signatures of Breast Cancer Aggressiveness
Diagnostic assessment of breast cancer biopsies remains reliant on resource-intensive histopathology and molecular profiling, which often lack real-time physiological readouts. Magnetic levitation (MagLev) enables label-free density profiling of single cells, yet its application to intact tissue biopsies has been precluded by size-dependent geometric artifacts and the absence of analytical frameworks for biopsy-scale samples. Here, we report the first application of MagLev to intact invasive breast carcinoma biopsies (200-600 μm) for biophysical profiling, generating multivariate biophysical signatures from 203 samples across 17 patients. We developed a physics-based size-correction algorithm (xmc) that isolates biological density from geometric artifact, and demonstrate that tissue viability is predicted not by average levitation height, but by spatial heterogeneity across replicate samples, reflecting the microenvironmental complexity of metabolically active tumors. Multivariate integration using Partial Least Squares (PLS) regression and Factor Analysis of Mixed Data (FAMD) identified nodal status (N) as the strongest biophysical predictor, suggesting that lymphatic dissemination capacity leaves a measurable signature in the primary tumor density profile. Unsupervised patient clustering in PLS-derived latent space recovered three clinically coherent subgroups aligned with molecular subtypes. This 30-minute, low-cost assay provides exploratory biophysical stratification complementary to existing diagnostics, particularly in resource-limited settings.
- 20 April 2026
- Published by SAGE Publications in Acta Radiologica
- Vol. 67 (6) , 517-527
T2* mapping of the patellofemoral joint at 3 T MRI: a repeatability and reproducibility study
Background: T2* mapping provides objective data on chondromalacia. However, its reliability in evaluating both patellar and trochlear cartilages has not yet been established. Purpose: To evaluate the repeatability and reproducibility of T2* mapping of patellar and trochlear cartilage in patients with patellofemoral maltracking, and to compare the reliability of axial and sagittal imaging planes for early detection of chondromalacia. Material and Methods: This retrospective study included 106 knees from patients aged 18–40 years presenting with anterior knee pain and MRI evidence of patellofemoral malalignment between January 2023 and August 2025. T2* mapping was performed on a 3 T MRI scanner using standardized protocols. Patellar and trochlear cartilage were manually divided into 12 compartments in axial and sagittal planes. Regions of interest were placed manually, and T2* relaxation times were measured independently by a musculoskeletal radiologist and an orthopedic surgeon. Intra- and inter-observer reliability were assessed using intraclass correlation coefficients (ICCs). Results: Most compartments demonstrated good-to-excellent intra- and inter-observer agreement (ICC ≥0.75), particularly in the upper and mid patellar cartilage zones. Axial-plane measurements consistently showed higher ICC values than sagittal-plane measurements. The lowest reliability was observed in lower cartilage zones and in sagittal-plane measurements. All results were statistically significant ( P <0.001). Conclusion: T2* mapping of the patellofemoral joint provides reliable measurements, with superior performance in the axial plane and in mid-upper zones. Routine use of axial-plane T2* mapping may facilitate early detection of chondromalacia in young adults with anterior knee pain, potentially improving clinical decision-making and preventing irreversible cartilage damage.
- 20 April 2026
- Published by SAGE Publications in Acta Radiologica
- Vol. 67 (6) , 517-527
T2* mapping of the patellofemoral joint at 3 T MRI: a repeatability and reproducibility study
Background: T2* mapping provides objective data on chondromalacia. However, its reliability in evaluating both patellar and trochlear cartilages has not yet been established. Purpose: To evaluate the repeatability and reproducibility of T2* mapping of patellar and trochlear cartilage in patients with patellofemoral maltracking, and to compare the reliability of axial and sagittal imaging planes for early detection of chondromalacia. Material and Methods: This retrospective study included 106 knees from patients aged 18–40 years presenting with anterior knee pain and MRI evidence of patellofemoral malalignment between January 2023 and August 2025. T2* mapping was performed on a 3 T MRI scanner using standardized protocols. Patellar and trochlear cartilage were manually divided into 12 compartments in axial and sagittal planes. Regions of interest were placed manually, and T2* relaxation times were measured independently by a musculoskeletal radiologist and an orthopedic surgeon. Intra- and inter-observer reliability were assessed using intraclass correlation coefficients (ICCs). Results: Most compartments demonstrated good-to-excellent intra- and inter-observer agreement (ICC ≥0.75), particularly in the upper and mid patellar cartilage zones. Axial-plane measurements consistently showed higher ICC values than sagittal-plane measurements. The lowest reliability was observed in lower cartilage zones and in sagittal-plane measurements. All results were statistically significant ( P <0.001). Conclusion: T2* mapping of the patellofemoral joint provides reliable measurements, with superior performance in the axial plane and in mid-upper zones. Routine use of axial-plane T2* mapping may facilitate early detection of chondromalacia in young adults with anterior knee pain, potentially improving clinical decision-making and preventing irreversible cartilage damage.
- Open Access
- 18 August 2025
- Published by Galenos Yayinevi in Diagnostic and Interventional Radiology
- Vol. 32 (4) , 492-499
Association of patellofemoral malalignment with early trochlear and patellar chondromalacia: a prospective T2* mapping study
Association of patellofemoral malalignment with early trochlear and patellar chondromalacia: a prospective T2* mapping study - Diagnostic and Interventional Radiology
- 18 August 2025
- Published by Galenos Yayinevi in Diagnostic and Interventional Radiology
- Vol. 32 (4) , 492-499
Open AccessAssociation of patellofemoral malalignment with early trochlear and patellar chondromalacia: a prospective T2* mapping study
Association of patellofemoral malalignment with early trochlear and patellar chondromalacia: a prospective T2* mapping study - Diagnostic and Interventional Radiology
- Open Access
- 17 July 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (14) , 1801
Can Mammography and Magnetic Resonance Imaging Predict the Preoperative Size and Nuclear Grade of Pure Ductal Carcinoma In Situ?
Background/Objectives: Thirty to fifty percent of ductal carcinoma in situ (DCIS) cases are high-grade and at risk of progressing to invasive carcinoma. The most important treatment-related risk factor for recurrence is the presence of residual DCIS. The aim of our study was to evaluate the relationship between size and imaging features on preoperative mammography and magnetic resonance imaging (MRI) and histopathological size and nuclear grade in patients with pure DCIS. Methods: Between 2015 and 2023, 90 patients who underwent surgery for DCIS, had no microinvasive/invasive component, and underwent a preoperative mammography and MRI were included in this study. Results: DCIS was detected in 91.1% of patients using mammography and 95.5% using MRI. Microcalcifications (MCs) were most common in mammography (85.4%). Thin pleomorphic and thin linear branching MCs were detected in 42% of high-grade DCIS, while amorphous (42%) MCs were most common in low-grade DCIS. In low-grade DCIS cases, a grouped distribution of MCs was observed most commonly (69%). There was a statistically significant difference between DCIS groups in terms of MC morphology and distribution (p = 0.043, p = 0.005, respectively). Diffusion restriction on MRI was associated with high-grade DCIS (p = 0.043). The tumor size was greater than the pathological size and correlated poorly with mammography and moderately with MRI. Conclusions: Compared to mammography, MRI is more effective in detecting and estimating the size of DCIS. Both methods overestimate tumor size compared to histopathological size. The nuclear grade is associated with a poor prognosis and local recurrence in DCIS.
- 17 July 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (14) , 1801
Open AccessCan Mammography and Magnetic Resonance Imaging Predict the Preoperative Size and Nuclear Grade of Pure Ductal Carcinoma In Situ?
Background/Objectives: Thirty to fifty percent of ductal carcinoma in situ (DCIS) cases are high-grade and at risk of progressing to invasive carcinoma. The most important treatment-related risk factor for recurrence is the presence of residual DCIS. The aim of our study was to evaluate the relationship between size and imaging features on preoperative mammography and magnetic resonance imaging (MRI) and histopathological size and nuclear grade in patients with pure DCIS. Methods: Between 2015 and 2023, 90 patients who underwent surgery for DCIS, had no microinvasive/invasive component, and underwent a preoperative mammography and MRI were included in this study. Results: DCIS was detected in 91.1% of patients using mammography and 95.5% using MRI. Microcalcifications (MCs) were most common in mammography (85.4%). Thin pleomorphic and thin linear branching MCs were detected in 42% of high-grade DCIS, while amorphous (42%) MCs were most common in low-grade DCIS. In low-grade DCIS cases, a grouped distribution of MCs was observed most commonly (69%). There was a statistically significant difference between DCIS groups in terms of MC morphology and distribution (p = 0.043, p = 0.005, respectively). Diffusion restriction on MRI was associated with high-grade DCIS (p = 0.043). The tumor size was greater than the pathological size and correlated poorly with mammography and moderately with MRI. Conclusions: Compared to mammography, MRI is more effective in detecting and estimating the size of DCIS. Both methods overestimate tumor size compared to histopathological size. The nuclear grade is associated with a poor prognosis and local recurrence in DCIS.
- 27 December 2023
- Published by Osmangazi Journal of Medicine in OSMANGAZİ JOURNAL OF MEDICINE
- Vol. 46 (2)
Memede Kollizyon Tümörü: Heterolog Komponentli Malign Filloides Tümör ve İnvaziv Meme Karsinomu, Spesifiye Edilemeyen Tip
- 27 December 2023
- Published by Osmangazi Journal of Medicine in OSMANGAZİ JOURNAL OF MEDICINE
- Vol. 46 (2)
Memede Kollizyon Tümörü: Heterolog Komponentli Malign Filloides Tümör ve İnvaziv Meme Karsinomu, Spesifiye Edilemeyen Tip
- 1 April 2023
- Published by Galenos Yayinevi in European Journal of Breast Health
- Vol. 19 (2) , 140-147
High-Grade Ductal Carcinoma In Situ of the Breast With Regressive Changes: Radiological and Clinicopathological Findings
Tumour regression is defined as continuity of changes leading to the elimination of a neoplastic population and is reflected as periductal fibrosis and intraductal tumour attenuation. The aim of this study was to describe the radiological and clinicopathological characteristics of high-grade breast ductal carcinoma in situ (DCIS) with regressive changes (RC). Thirty-two cases of high-grade DCIS with RC on biopsy specimens followed by excision were included. The mammographic, ultrasonographic (US), and magnetic resonance imaging (MRI) findings of cases were retrospectively reviewed according to the breast imaging reporting and data system (BI-RADS) lexicon. Clinical and histopathological findings [comedonecrosis, estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2) status and Ki-67 proliferation index] were recorded. The rate of upgrade to invasive cancer after surgical excision and lymph node involvement were evaluated. The most common mammographic finding was microcalcifications alone (68.8%). The most frequently seen findings on US were microcalcifications only (21.9%), followed by microcalcifications and hypoechoic area (18.7%). On MRI, most lesions presented as clumped non-mass enhancement with segmental distribution. ER/PR negativity (53.1%, 65.6%), HER2 positivity (56.3%) and high Ki-67 (62.5%), which are known to be associated with more aggressive behavior, were found to be proportionally higher. The rate of upgrade to invasive cancer was 21.8%. DCIS with RC lesions present most often as microcalcifications alone on both mammography and US. MRI features are not distinguishable from those of other DCIS lesions. DCIS with RC lesions show biomarker status reflecting more aggressive behavior and high upgrade rate to invasive cancer.
- 1 April 2023
- Published by Galenos Yayinevi in European Journal of Breast Health
- Vol. 19 (2) , 140-147
High-Grade Ductal Carcinoma In Situ of the Breast With Regressive Changes: Radiological and Clinicopathological Findings
Tumour regression is defined as continuity of changes leading to the elimination of a neoplastic population and is reflected as periductal fibrosis and intraductal tumour attenuation. The aim of this study was to describe the radiological and clinicopathological characteristics of high-grade breast ductal carcinoma in situ (DCIS) with regressive changes (RC). Thirty-two cases of high-grade DCIS with RC on biopsy specimens followed by excision were included. The mammographic, ultrasonographic (US), and magnetic resonance imaging (MRI) findings of cases were retrospectively reviewed according to the breast imaging reporting and data system (BI-RADS) lexicon. Clinical and histopathological findings [comedonecrosis, estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2) status and Ki-67 proliferation index] were recorded. The rate of upgrade to invasive cancer after surgical excision and lymph node involvement were evaluated. The most common mammographic finding was microcalcifications alone (68.8%). The most frequently seen findings on US were microcalcifications only (21.9%), followed by microcalcifications and hypoechoic area (18.7%). On MRI, most lesions presented as clumped non-mass enhancement with segmental distribution. ER/PR negativity (53.1%, 65.6%), HER2 positivity (56.3%) and high Ki-67 (62.5%), which are known to be associated with more aggressive behavior, were found to be proportionally higher. The rate of upgrade to invasive cancer was 21.8%. DCIS with RC lesions present most often as microcalcifications alone on both mammography and US. MRI features are not distinguishable from those of other DCIS lesions. DCIS with RC lesions show biomarker status reflecting more aggressive behavior and high upgrade rate to invasive cancer.
- Open Access
- 12 January 2023
- Published by Galenos Yayinevi in Diagnostic and Interventional Radiology
- Vol. 29 (1) , 53-58
Radiological and clinicopathological findings of breast cancer during the COVID-19 pandemic: a comparative study with the pre-pandemic era
PURPOSE The diagnosis and surgical treatment delays that occurred during the coronavirus disease-2019(COVID-19) pandemic may have affected breast cancer presentation. This study aimed to determine whether there was a difference in the clinicopathological characteristics of breast cancers during the pandemic by comparing them with similar cases from the previous year. The study also aimed to determine the radiological findings of breast cancers during the pandemic. METHODS A retrospective review was made of patients who underwent surgery for breast cancer between March 11, 2020, and December 11, 2020 (the pandemic group). These patients were compared with similar patients from the previous year (the pre-pandemic group). The postoperative histopathology results of both groups were compared, and the preoperative radiological findings of the pandemic group were defined. RESULTS There were 71 patients in the pandemic group and 219 patients in the pre-pandemic group. The tumor size was significantly greater, lymph node involvement was more frequent, and waiting time for surgery was longer in the pandemic group (P < 0.001, P = 0.044, P = 0.001, respectively). There was no significant difference between the groups in respect of in situ/invasive tumor distribution, histological type and histological grade of tumor, the presence of lymphovascular/perineural invasion, multifocal/multicentric focus, and Breast Imaging Reporting and Data System Classification (P > 0.15). The radiologic findings of breast cancer during the pandemic typically showed characteristics of malignancy. CONCLUSION Patients diagnosed with breast cancer during the COVID-19 pandemic had larger tumor sizes, more frequent lymph node involvement and longer waiting time for surgical treatment. Screening programs should be continued as soon as possible by taking necessary precautions.
- 12 January 2023
- Published by Galenos Yayinevi in Diagnostic and Interventional Radiology
- Vol. 29 (1) , 53-58
Open AccessRadiological and clinicopathological findings of breast cancer during the COVID-19 pandemic: a comparative study with the pre-pandemic era
PURPOSE The diagnosis and surgical treatment delays that occurred during the coronavirus disease-2019(COVID-19) pandemic may have affected breast cancer presentation. This study aimed to determine whether there was a difference in the clinicopathological characteristics of breast cancers during the pandemic by comparing them with similar cases from the previous year. The study also aimed to determine the radiological findings of breast cancers during the pandemic. METHODS A retrospective review was made of patients who underwent surgery for breast cancer between March 11, 2020, and December 11, 2020 (the pandemic group). These patients were compared with similar patients from the previous year (the pre-pandemic group). The postoperative histopathology results of both groups were compared, and the preoperative radiological findings of the pandemic group were defined. RESULTS There were 71 patients in the pandemic group and 219 patients in the pre-pandemic group. The tumor size was significantly greater, lymph node involvement was more frequent, and waiting time for surgery was longer in the pandemic group (P < 0.001, P = 0.044, P = 0.001, respectively). There was no significant difference between the groups in respect of in situ/invasive tumor distribution, histological type and histological grade of tumor, the presence of lymphovascular/perineural invasion, multifocal/multicentric focus, and Breast Imaging Reporting and Data System Classification (P > 0.15). The radiologic findings of breast cancer during the pandemic typically showed characteristics of malignancy. CONCLUSION Patients diagnosed with breast cancer during the COVID-19 pandemic had larger tumor sizes, more frequent lymph node involvement and longer waiting time for surgical treatment. Screening programs should be continued as soon as possible by taking necessary precautions.
- 1 September 2021
- Published by Elsevier in Hand Surgery and Rehabilitation
- Vol. 40 (4) , 427-432
Radial shortening osteotomy vs partial capitate shortening osteotomy in Kienböck’s disease: Medium-term radiological and clinical results
- 1 September 2021
- Published by Elsevier in Hand Surgery and Rehabilitation
- Vol. 40 (4) , 427-432
Radial shortening osteotomy vs partial capitate shortening osteotomy in Kienböck’s disease: Medium-term radiological and clinical results