Najoua Lassoued
Research Subjects
Research Interests
Najoua Lassoued has a total of 54 co-authors. Here are the top co-authors by number of shared publications:
K. Mighri
Service d’ORL, CHU Tahar Sfar Mahdia, Mahdia, Tunisie
4 shared publicationsE. Chebil
Service d’ORL, CHU Tahar Sfar Mahdia, Mahdia, Tunisie
3 shared publicationsF. Mnif
Service d’endocrinologie, CHU Hédi Chaker, Sfax, Tunisie
3 shared publicationsM. Abid
Service d’endocrinologie, CHU Hédi-Chaker, Sfax, Tunisie
3 shared publicationsM. Hasnaoui
Service d’ORL, CHU Tahar Sfar Mahdia, Mahdia, Tunisie
3 shared publicationsM. Masmoudi
Service d’ORL, CHU Tahar Sfar Mahdia, Mahdia, Tunisie
3 shared publications
Latest Publications
- Open Access
Fine-tuned Medical Vision–Language Model (MedGemma 1.5 4B) with and without Structured Radiologist Descriptors versus EU-TIRADS for Malignancy Risk Stratification of Thyroid Nodules at Ultrasound: A Tunisian Multicenter Diagnostic Accuracy Study
Purpose: To evaluate a medical vision–language foundation model fine-tuned on thyroid ultrasound images, with or without the radiologist’s structured descriptors, against expert-rated EU-TIRADS for the classification of thyroid nodules as benign or malignant. Materials and Methods: This retrospective study included nodules that underwent ultrasound-guided fine-needle aspiration (FNA) between January 2019 and August 2025 in two Tunisian centers, with consecutive recruitment, together with a selection of images from an international public dataset. Of 1,580 nodules, 1,456 (272 malignant; 1,280 patients) with an interpretable cytological or histological reference standard were allocated at patient level to training (n = 942), validation (n = 258) and a test set sealed before training (n = 256; 51 malignant). MedGemma 1.5 4B was fine-tuned with low-rank adaptation under four input configurations: the image with nodule biometry, with or without age and sex (A, B); A with chain-of-thought supervision (C); and A plus the eleven qualitative descriptors of the radiologist’s structured report (D). The EU-TIRADS category was never a model input. Areas under the receiver operating characteristic curve (AUROCs) were compared with DeLong’s test and a prespecified non-inferiority margin of 0.05; EU-TIRADS ≥ 4 was the prespecified comparator, with ≥ 5 as a sensitivity analysis. Results: Config D reached an AUROC of 0.953 (95% CI 0.909–0.976), versus 0.948 (0.894–0.972) for EU-TIRADS (difference +0.004, 95% CI −0.016 to +0.025), which met the non-inferiority criterion; Configs A–C, without the qualitative descriptors, reached 0.883–0.892 (Holm-adjusted p = 0.045 versus EU-TIRADS). At ≥ 4, EU-TIRADS had a sensitivity of 98.0% and produced 94 false positives among 205 benign nodules; at a threshold fixed on validation data, Config D reclassified 72 of them as negative while missing two additional cancers. Against EU-TIRADS ≥ 5 (sensitivity 92.2%, specificity 95.1%), Config D at its native threshold (88.2%, 94.1%) did not classify more nodules correctly (McNemar p = 0.22), and at its two other thresholds it classified fewer nodules correctly (p = 0.004 and 0.007). Raw probabilities were miscalibrated (expected calibration error 0.117 for Config D), and Platt scaling within the test set reduced this error to 0.019–0.040. Conclusion: Given the ultrasound image together with the radiologist’s structured descriptors, the fine-tuned model matched expert EU-TIRADS discrimination and, against the ≥ 4 threshold, reclassified most false positives as negative, a gain concentrated in category 4 and based on only three cancers in category 4. It is best positioned as a report-standardizing second reader in the EU-TIRADS 4 gray zone, pending prospective external validation. Summary Statement: A fine-tuned medical vision–language model given the ultrasound image and the radiologist’s structured report matched expert EU-TIRADS discrimination of thyroid nodules and would have deferred most aspirations of benign category 4 nodules, at the cost of one or two missed category 4 cancers depending on the threshold, based on only three cancers in category 4. Key Points: On a test set of 256 thyroid nodules sealed before training, the model given the ultrasound image and the radiologist’s 16-field structured report reached an area under the receiver operating characteristic curve of 0.953, non-inferior to expert EU-TIRADS (0.948). Against EU-TIRADS 4 or higher, the model reclassified as negative 72 of the 94 false-positive nodules at a threshold fixed on validation data, at the cost of two additional missed cancers. Adding the radiologist’s qualitative descriptors to the image raised the area under the curve from 0.892 (Config A) to 0.953, and Platt scaling reduced the calibration error of the model from 0.117 to 0.019–0.040.
Open AccessFine-tuned Medical Vision–Language Model (MedGemma 1.5 4B) with and without Structured Radiologist Descriptors versus EU-TIRADS for Malignancy Risk Stratification of Thyroid Nodules at Ultrasound: A Tunisian Multicenter Diagnostic Accuracy Study
Purpose: To evaluate a medical vision–language foundation model fine-tuned on thyroid ultrasound images, with or without the radiologist’s structured descriptors, against expert-rated EU-TIRADS for the classification of thyroid nodules as benign or malignant. Materials and Methods: This retrospective study included nodules that underwent ultrasound-guided fine-needle aspiration (FNA) between January 2019 and August 2025 in two Tunisian centers, with consecutive recruitment, together with a selection of images from an international public dataset. Of 1,580 nodules, 1,456 (272 malignant; 1,280 patients) with an interpretable cytological or histological reference standard were allocated at patient level to training (n = 942), validation (n = 258) and a test set sealed before training (n = 256; 51 malignant). MedGemma 1.5 4B was fine-tuned with low-rank adaptation under four input configurations: the image with nodule biometry, with or without age and sex (A, B); A with chain-of-thought supervision (C); and A plus the eleven qualitative descriptors of the radiologist’s structured report (D). The EU-TIRADS category was never a model input. Areas under the receiver operating characteristic curve (AUROCs) were compared with DeLong’s test and a prespecified non-inferiority margin of 0.05; EU-TIRADS ≥ 4 was the prespecified comparator, with ≥ 5 as a sensitivity analysis. Results: Config D reached an AUROC of 0.953 (95% CI 0.909–0.976), versus 0.948 (0.894–0.972) for EU-TIRADS (difference +0.004, 95% CI −0.016 to +0.025), which met the non-inferiority criterion; Configs A–C, without the qualitative descriptors, reached 0.883–0.892 (Holm-adjusted p = 0.045 versus EU-TIRADS). At ≥ 4, EU-TIRADS had a sensitivity of 98.0% and produced 94 false positives among 205 benign nodules; at a threshold fixed on validation data, Config D reclassified 72 of them as negative while missing two additional cancers. Against EU-TIRADS ≥ 5 (sensitivity 92.2%, specificity 95.1%), Config D at its native threshold (88.2%, 94.1%) did not classify more nodules correctly (McNemar p = 0.22), and at its two other thresholds it classified fewer nodules correctly (p = 0.004 and 0.007). Raw probabilities were miscalibrated (expected calibration error 0.117 for Config D), and Platt scaling within the test set reduced this error to 0.019–0.040. Conclusion: Given the ultrasound image together with the radiologist’s structured descriptors, the fine-tuned model matched expert EU-TIRADS discrimination and, against the ≥ 4 threshold, reclassified most false positives as negative, a gain concentrated in category 4 and based on only three cancers in category 4. It is best positioned as a report-standardizing second reader in the EU-TIRADS 4 gray zone, pending prospective external validation. Summary Statement: A fine-tuned medical vision–language model given the ultrasound image and the radiologist’s structured report matched expert EU-TIRADS discrimination of thyroid nodules and would have deferred most aspirations of benign category 4 nodules, at the cost of one or two missed category 4 cancers depending on the threshold, based on only three cancers in category 4. Key Points: On a test set of 256 thyroid nodules sealed before training, the model given the ultrasound image and the radiologist’s 16-field structured report reached an area under the receiver operating characteristic curve of 0.953, non-inferior to expert EU-TIRADS (0.948). Against EU-TIRADS 4 or higher, the model reclassified as negative 72 of the 94 false-positive nodules at a threshold fixed on validation data, at the cost of two additional missed cancers. Adding the radiologist’s qualitative descriptors to the image raised the area under the curve from 0.892 (Config A) to 0.953, and Platt scaling reduced the calibration error of the model from 0.117 to 0.019–0.040.
- Open Access
From meal text to carbohydrate to postprandial glucose rise: QLoRA fine-tuning of the Qwen3-4B language model and gradient-boosting prediction in type 2 diabetes
Background: In people with type 2 diabetes (T2D), postprandial glucose weighs on glycaemic control, and carbohydrate counting is difficult. Patients can describe their meal but rarely quantify it. Objective: To assess (1) whether a small open large language model (LLM), fine-tuned on public data, estimates the carbohydrate content of a meal described in words, and (2) what this carbohydrate adds to the prediction of the postprandial glucose rise in two public cohorts. Methods: Step 1: the Qwen3-4B LLM was fine-tuned with QLoRA on Nutrition5k dishes and compared with an automatically matched USDA food composition table and with Qwen3-4B and Qwen3-8B used without training, on test dishes (internal control) and on 100 real meals of ShanghaiT2DM patients, whose reference carbohydrate was computed from the recorded quantities. Step 2: the maximum glucose rise between 0 and 120 minutes (ΔGmax) was predicted by personalised gradient boosting (M2), compared with a spline linear regression and with the Chronos-2 foundation model, using only information available at meal start and person-level cross-validation, in ShanghaiT2DM (people with T2D recruited in two Shanghai hospitals; carbohydrate estimated by the LLM) and CGMacros (free-living adults in Santa Barbara, California; carbohydrate entered in the dataset, no text). Step 3: the model was transferred between datasets. 95% confidence intervals (CI) were obtained by cluster BCa bootstrap (2,000 resamples). Results: On the 100 real meals, the carbohydrate mean absolute error (MAE) of the fine-tuned model was 12.2 [9.9, 15.2] g, almost half that of the table (22.6 g; paired difference -10.5 [-16.8, -6.0] g) and nearly a third of that of the same model without fine-tuning (34.3 g), with a bias of -10.3 g; on test dishes it was 2.2 [2.0, 2.5] g. Personalised gradient boosting (M2) reached a correlation between predicted and observed rises of r = 0.63 [0.55, 0.69] in CGMacros (1,286 meals, 45 adults) and 0.49 [0.42, 0.54] in ShanghaiT2DM (2,339 meals, 99 people with T2D), about twice the linear reference (r = 0.32 and 0.23); for a rise above 50 mg/dL, its area under the ROC curve (AUROC) was 0.80 and 0.74, versus 0.64 and 0.68 with pre-meal glucose alone. Adding carbohydrate to pre-meal glucose changed r by 0.095 [0.068, 0.128] in CGMacros and by 0.004 [-0.010, 0.017] in ShanghaiT2DM; difference between datasets: 0.091 [0.060, 0.126]. Applied from one dataset to the other, the model lost correlation (Δr = -0.11 [-0.17, -0.05] towards ShanghaiT2DM and -0.12 [-0.19, -0.04] towards CGMacros). The full “text → LLM → glucose” chain was evaluated in ShanghaiT2DM only. Conclusion: An open 4-billion-parameter LLM, fine-tuned with QLoRA on public data and usable offline, almost halves the error of carbohydrate counting for real meals described in words compared with an automatically matched table, with an underestimation of large meals. Personalised gradient boosting, based on pre-meal glucose and the person’s usual response, doubles the correlation of a linear reference for predicting the glucose rise in both cohorts. Carbohydrate improved this prediction in CGMacros but not in ShanghaiT2DM; the two datasets differ in many respects, and the cause of this gap is not established.
Open AccessFrom meal text to carbohydrate to postprandial glucose rise: QLoRA fine-tuning of the Qwen3-4B language model and gradient-boosting prediction in type 2 diabetes
Background: In people with type 2 diabetes (T2D), postprandial glucose weighs on glycaemic control, and carbohydrate counting is difficult. Patients can describe their meal but rarely quantify it. Objective: To assess (1) whether a small open large language model (LLM), fine-tuned on public data, estimates the carbohydrate content of a meal described in words, and (2) what this carbohydrate adds to the prediction of the postprandial glucose rise in two public cohorts. Methods: Step 1: the Qwen3-4B LLM was fine-tuned with QLoRA on Nutrition5k dishes and compared with an automatically matched USDA food composition table and with Qwen3-4B and Qwen3-8B used without training, on test dishes (internal control) and on 100 real meals of ShanghaiT2DM patients, whose reference carbohydrate was computed from the recorded quantities. Step 2: the maximum glucose rise between 0 and 120 minutes (ΔGmax) was predicted by personalised gradient boosting (M2), compared with a spline linear regression and with the Chronos-2 foundation model, using only information available at meal start and person-level cross-validation, in ShanghaiT2DM (people with T2D recruited in two Shanghai hospitals; carbohydrate estimated by the LLM) and CGMacros (free-living adults in Santa Barbara, California; carbohydrate entered in the dataset, no text). Step 3: the model was transferred between datasets. 95% confidence intervals (CI) were obtained by cluster BCa bootstrap (2,000 resamples). Results: On the 100 real meals, the carbohydrate mean absolute error (MAE) of the fine-tuned model was 12.2 [9.9, 15.2] g, almost half that of the table (22.6 g; paired difference -10.5 [-16.8, -6.0] g) and nearly a third of that of the same model without fine-tuning (34.3 g), with a bias of -10.3 g; on test dishes it was 2.2 [2.0, 2.5] g. Personalised gradient boosting (M2) reached a correlation between predicted and observed rises of r = 0.63 [0.55, 0.69] in CGMacros (1,286 meals, 45 adults) and 0.49 [0.42, 0.54] in ShanghaiT2DM (2,339 meals, 99 people with T2D), about twice the linear reference (r = 0.32 and 0.23); for a rise above 50 mg/dL, its area under the ROC curve (AUROC) was 0.80 and 0.74, versus 0.64 and 0.68 with pre-meal glucose alone. Adding carbohydrate to pre-meal glucose changed r by 0.095 [0.068, 0.128] in CGMacros and by 0.004 [-0.010, 0.017] in ShanghaiT2DM; difference between datasets: 0.091 [0.060, 0.126]. Applied from one dataset to the other, the model lost correlation (Δr = -0.11 [-0.17, -0.05] towards ShanghaiT2DM and -0.12 [-0.19, -0.04] towards CGMacros). The full “text → LLM → glucose” chain was evaluated in ShanghaiT2DM only. Conclusion: An open 4-billion-parameter LLM, fine-tuned with QLoRA on public data and usable offline, almost halves the error of carbohydrate counting for real meals described in words compared with an automatically matched table, with an underestimation of large meals. Personalised gradient boosting, based on pre-meal glucose and the person’s usual response, doubles the correlation of a linear reference for predicting the glucose rise in both cohorts. Carbohydrate improved this prediction in CGMacros but not in ShanghaiT2DM; the two datasets differ in many respects, and the cause of this gap is not established.
- Open Access
- 1 April 2023
- Published by Bioscientifica in Endocrinology, Diabetes & Metabolism Case Reports
- Vol. 2023 (2)
Acute adrenal crisis precipitated by thyroid storm in a patient with undiagnosed autoimmune polyglandular syndrome type 2
Summary: Autoimmune polyglandular syndrome (APS) type 2 is characterized by the presence of Addison’s disease (AD) along with autoimmune thyroid disease and/or type 1 diabetes. APS type 2 is known as Schmidt’s syndrome when autoimmune adrenal insufficiency is associated with chronic lymphocytic thyroiditis. We report a very rare case of a 28-year-old female patient who had Schmidt’s syndrome revealed by a thyroid storm (TS) concomitant with an acute adrenal crisis. The onset of AD resulted in a surgical emergency. The patient presented with cardiogenic shock and an acute abdomen. The precipitation factor was Hashitoxicosis presented as TS. This life-threatening condition was successfully reversed with aggressive medical therapy based on antithyroid drugs and intravenous glucocorticoids. This hyperthyroid phase lasted for a period of 8 months. The patient eventually developed hypothyroidism, suggesting that Hashimoto's thyroiditis was the most likely diagnosis. She was started on levothyroxine replacement therapy and remained euthyroid on levothyroxine. The case we describe had several diagnostic pitfalls that are discussed both at the start as well as during the evolution. Learning points: Autoimmune diseases can appear concomitantly or succeed each other over time. The clinician must be vigilant to detect these diseases in time in order to avoid a misdiagnosis of a life-threatening emergency such as adrenal insufficiency or thyroid storm. Thyroid storm is an uncommon but life-threatening manifestation of hyperthyroidism. Diagnosis is dependent on clinical symptoms, and no specific laboratory tests are available. Glucocorticoids should be used in the treatment of thyroid storm because they have an inhibitory effect on peripheral conversion of T4 to T3. In patients who have severe thyrotoxicosis, especially in conjunction with hypotension, treatment with glucocorticoids has become standard practice because of the possibility of relative adrenal insufficiency or the possibility of undiagnosed Addison’s disease. The differential diagnosis of hyperthyroidism can be challenging. Graves’ disease can be discussed in view of the severity of the clinical presentation and the prolonged duration of the hyperthyroid phase. Hashitoxicosis is the initial hyperthyroid phase in chronic autoimmune thyroiditis. The hyperthyroid phase is always followed by definitive resolution, with persistent euthyroidism and no hyperthyroid relapses. Synthetic antithyroid drugs may be prescribed during the hyperthyroid phase of Hashimoto thyroiditis if the clinical presentation is severe and the duration of the hyperthyroid phase is prolonged.
- 1 April 2023
- Published by Bioscientifica in Endocrinology, Diabetes & Metabolism Case Reports
- Vol. 2023 (2)
Open AccessAcute adrenal crisis precipitated by thyroid storm in a patient with undiagnosed autoimmune polyglandular syndrome type 2
Summary: Autoimmune polyglandular syndrome (APS) type 2 is characterized by the presence of Addison’s disease (AD) along with autoimmune thyroid disease and/or type 1 diabetes. APS type 2 is known as Schmidt’s syndrome when autoimmune adrenal insufficiency is associated with chronic lymphocytic thyroiditis. We report a very rare case of a 28-year-old female patient who had Schmidt’s syndrome revealed by a thyroid storm (TS) concomitant with an acute adrenal crisis. The onset of AD resulted in a surgical emergency. The patient presented with cardiogenic shock and an acute abdomen. The precipitation factor was Hashitoxicosis presented as TS. This life-threatening condition was successfully reversed with aggressive medical therapy based on antithyroid drugs and intravenous glucocorticoids. This hyperthyroid phase lasted for a period of 8 months. The patient eventually developed hypothyroidism, suggesting that Hashimoto's thyroiditis was the most likely diagnosis. She was started on levothyroxine replacement therapy and remained euthyroid on levothyroxine. The case we describe had several diagnostic pitfalls that are discussed both at the start as well as during the evolution. Learning points: Autoimmune diseases can appear concomitantly or succeed each other over time. The clinician must be vigilant to detect these diseases in time in order to avoid a misdiagnosis of a life-threatening emergency such as adrenal insufficiency or thyroid storm. Thyroid storm is an uncommon but life-threatening manifestation of hyperthyroidism. Diagnosis is dependent on clinical symptoms, and no specific laboratory tests are available. Glucocorticoids should be used in the treatment of thyroid storm because they have an inhibitory effect on peripheral conversion of T4 to T3. In patients who have severe thyrotoxicosis, especially in conjunction with hypotension, treatment with glucocorticoids has become standard practice because of the possibility of relative adrenal insufficiency or the possibility of undiagnosed Addison’s disease. The differential diagnosis of hyperthyroidism can be challenging. Graves’ disease can be discussed in view of the severity of the clinical presentation and the prolonged duration of the hyperthyroid phase. Hashitoxicosis is the initial hyperthyroid phase in chronic autoimmune thyroiditis. The hyperthyroid phase is always followed by definitive resolution, with persistent euthyroidism and no hyperthyroid relapses. Synthetic antithyroid drugs may be prescribed during the hyperthyroid phase of Hashimoto thyroiditis if the clinical presentation is severe and the duration of the hyperthyroid phase is prolonged.
- 1 October 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 414
Extension thyroïdienne dans les cancers du larynx
- 1 October 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 414
Extension thyroïdienne dans les cancers du larynx
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 457
Un syndrome de cellules de Sertoli seulement secondaire à une microdélétion du chromosome Y
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 457
Un syndrome de cellules de Sertoli seulement secondaire à une microdélétion du chromosome Y
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 313
Kyste parathyroïdien symptomatique et non fonctionnel chez un adulte
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 313
Kyste parathyroïdien symptomatique et non fonctionnel chez un adulte
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 301-302
Facteurs échographiques prédictifs de malignité dans les cancers différenciés de la thyroïde
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 301-302
Facteurs échographiques prédictifs de malignité dans les cancers différenciés de la thyroïde
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 302
Les tumeurs oncocytaires de la thyroïde
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 302
Les tumeurs oncocytaires de la thyroïde
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 383-384
Une hyperprolactinémie transitoire après prise de sulpiride révélant un macro adénome hypophysaire non sécrétant chez une jeune patiente
- 24 September 2021
- Published by Elsevier in Annales d'Endocrinologie
- Vol. 82 (5) , 383-384
Une hyperprolactinémie transitoire après prise de sulpiride révélant un macro adénome hypophysaire non sécrétant chez une jeune patiente