Search

Background Injectable poly-D,L-lactic acid (PDLLA), under the brand name of AestheFill (REGEN Biotech, Inc. Seoul, South Korea), is a biocompatible, biodegradable, and biostimulatory product utilized to correct soft tissue volume loss. Its efficacy and safety have not been fully studied in a large cohort. Objectives The aim of the study was to evaluate the efficacy and safety of a novel dermal filler injectable, poly-D,L-lactic acid. Methods This was an evaluator-blinded, multicenter, randomized controlled trial to compare the efficacy and safety of PDLLA vs hyaluronic acid in the correction of the nasolabial fold. Two hundred and sixty patients with moderate to severe nasolabial folds were enrolled and randomized to the treatment group (PDLLA) or control group (hyaluronic acid). Each patient received a PDLLA or hyaluronic acid injection for nasolabial fold augmentation and was followed for 52 weeks. The Wrinkle Severity Rating Scale (WSRS) and Global Aesthetic Improvement Scale (GAIS) were applied to evaluate topical nasolabial fold augmentation and overall improvement, respectively. Results At 24 weeks, 67.6% of patients in the PDLLA group had an at least 1-grade improvement on the WSRS, compared to 60.9% of patients in the control group with an at least 1-grade improvement on the WSRS (P <.05). At each visit, the PDLLA group showed more improvement on the baseline WSRS than the control group. PDLLA was safe and well tolerated, with no severe adverse events. Conclusions PDLLA shows noninferior efficacy in correcting nasolabial folds compared to hyaluronic acid. Level of Evidence: 2 (Therapeutic) [ABSTRACT FROM AUTHOR]

Background: The feasibility of diagnosing malnutrition using facial features has been validated. A tool to integrate all facial features associated with malnutrition for disease screening is still demanded. This work aims to develop and evaluate a deep learning (DL) framework to accurately determine malnutrition based on a 3D facial points cloud. Methods: A group of 482 patients were studied in this perspective work. The 3D facial data were obtained using a 3D camera and represented as a 3D facial points cloud. A DL model, PointNet++, was trained and evaluated using the points cloud as inputs and classified the malnutrition states. The performance was evaluated with the area under the receiver operating characteristic curve, accuracy, specificity, sensitivity, and F1 score. Results: Among the 482 patients, 150 patients (31.1%) were diagnosed as having moderate malnutrition and 54 patients (11.2%) as having severe malnutrition. The DL model achieved the performance with an area under the receiver operating characteristic curve of 0.7240 ± 0.0416. Conclusion: The DL model achieved encouraging performance in accurately classifying nutrition states based on a points cloud of 3D facial information of patients with malnutrition. [ABSTRACT FROM AUTHOR]

Background: Three-dimensional (3D) stereophotogrammetry has been widely used in anthropometry for both medical and esthetic purposes. However, no studies have assessed its reliability on measuring the lower eyelid by 3D imaging. This study aimed to establish a standardized 3D anthropometric protocol for lower eyelid region and validate its reliability. Methods: Fifty-eight participants (116 eyes) were recruited with mean age of 39.14 ± 11.25 years. Two sets of VECTRA 3D images were taken for each subject, and each set of images was individually measured twice by two raters. Twenty-seven landmarks were identified in the lower eyelid region, and then 19 linear, 4 curvilinear, 7 angular and 2 areal metrics were assessed for intrarater, interrater and intramethod reliability. Results: High reliability was found in this 3D imaging-based lower eyelid anthropometry. The mean absolute difference within 2 intrarater measurement were 0.22 and 0.08 units, the technical errors measurement were 0.31 and 0.15 units, the relative errors of measurement were 0.90% and 0.31%, the relative technical errors of measurement were 1.53% and 0.64%, and the intra-group correlation coefficient was 0.99 and 0.99. The results for interrater measurement were 0.53 units, 0.59 units, 2.94%, 3.41% and 0.96, and for intramethod measurement were 0.71 units, 0.77 units, 4.12%, 4.05% and 0.95 units, respectively. Conclusion: 3D stereophotogrammetry is reliable for lower eyelid anthropometry. The standardized protocol can be further applied for many purposes such as lower eyelid aging evaluation, surgical related assessment and periocular rejuvenation plan. Level of Evidence IV: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266. [ABSTRACT FROM AUTHOR]

Background Botulinum toxin is administered to paralyze the gastrocnemius muscle and reduce its size, thereby improving the calf contour and reducing the leg circumference of the calf. Objectives In pursuit of better efficacy, we designed a new injection protocol that targeted both the gastrocnemius and soleus. An algorithm of botulinum toxin injection for calf contouring was proposed based on the results of the study and ultrasonographic data. Methods A prospective, self-controlled, double-blind study was conducted. The gastronemius muscle (GM) group (n = 17) included the patients whose gastrocnemius muscles were treated, and the GM + soleus muscle (SM) group (n = 17) included the patients who had both the gastrocnemius and soleus treated. Parameters including the maximum leg circumference and the subcutaneous fat, gastrocnemius muscle, and soleus muscle thicknesses were collected before and after injection. Results Both GM and GM + SM injection helped improve calf contour. Although the thickness of the gastrocnemius muscle was reduced, reducing the leg circumference, the subcutaneous fat and soleus muscle had compensatory thickening after injection, especially over the long term, which may affect the efficacy. Compared with the GM group, the GM + SM group effectively reduced the tendency of the soleus muscle to thicken. Calf muscle thickness could be roughly estimated by constructing a fitting equation and measuring height, weight, and leg circumference. Conclusions Two injection methods achieved the effect of improving calf contour and reducing the leg circumference with equivalent patient satisfaction. The GM + SM injection group did not show any obvious extra clinical benefit when compared with the GM injection group. Level of Evidence: 2 [ABSTRACT FROM AUTHOR]

Background Pretarsal rolls are a crucial aesthetic feature. Despite numerous studies conducted on periorbital aesthetics, there remains a dearth of research on the ideal pretarsal rolls. Objectives This study aimed to investigate 4 aesthetic characteristics of ideal pretarsal rolls: presence, width, proportion, and morphology. Methods Respondents (385, including 80 aesthetic specialists) were presented with 7 series of images of pretarsal rolls, including 2 series of pretarsal roll existence, 2 series of varying widths, 2 series of different ratios between pretarsal roll and palpebral fissure height, and 1 series about morphology. Participants were asked to rank each image within a given series from most attractive to least attractive. The rankings were then analyzed according to population demographics. Results The majority of respondents deemed images with pretarsal rolls to be more aesthetically pleasing (P <.001) than those without such features. Additionally, pretarsal rolls with a width of 5 mm (P <.001), a ratio to palpebral fissure height of 0.5:1 (P <.001), and a crescent shape (P <.001) were perceived as the most attractive. Moreover, individuals aged 40 or younger exhibited a significant preference for images with pretarsal rolls compared to older groups (P <.001), and females displayed a greater inclination toward a medium ratio of pretarsal roll to palpebral fissure height than males did (0.5:1 P =.003, 0.618:1 P <.001). Conclusions This study validates the optimal characteristics of pretarsal rolls, which provides insight into pretarsal roll aesthetics and holds significant implications for surgical planning and effect assessment in future periocular procedures involving pretarsal rolls. Level of Evidence: 3 [ABSTRACT FROM AUTHOR]

Background: Existing evaluation tools of tear trough deformity are based on subjective impression of clinicians. More accurate quantitative assessment methods are needed. This study aimed to propose a quantitative three-dimensional assessment method for the tear trough deformity in comparison with the Barton's grading system and apply it to the efficacy evaluation of orbital septum fat transposition. Methods: 117 healthy Chinese adults (234 eyes) were enrolled and divided into four groups according to the Barton's grading system. Three-dimensional facial images were captured using Vectra H1 handheld camera. 6 anthropometric landmarks were identified on each eye and 8 linear measurements were generated accordingly. Intra-observer reliability was determined and measurements were compared between groups. Pre- and post-operative three-dimensional measurements were compared in 19 patients who received lower blepharoplasty with orbital septum fat transposition. Results: The severity of tear trough was positively correlated with age (P < 0.001) but not BMI (P = 0.145) or gender (P = 0.280). Intra-rater reliability of the 8 linear measurements was excellent except for the vertical distance between the palpebrale inferioris margin and the tear trough below the pupillary center. Intergroup comparison showed that the horizontal distance between the lateral end of tear trough (P < 0.001) and medial canthus and the sagittal vector from tear trough point toward eyelid bag point (P = 0.009) increased with grade, while the vertical distance from mid-pupil to palpebrale inferioris margin decreased gradually (P = 0.001). Orbital septum fat transposition significantly improved the tear trough deformity as assessed by these three-dimensional measurements. Conclusions: We demonstrated a novel quantitative evaluation method of the tear trough deformity using three-dimensional stereophotogrammetry and proved it to be valid and reliable. It showed good value of clinical application and might help with periorbital rejuvenation planning and outcome assessment. Level of Evidence IV: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266. [ABSTRACT FROM AUTHOR]

Background: The non-invasive three-dimensional (3D) stereophotogrammetry is widely used in anthropometry for medical purpose. Yet, few studies have assessed its reliability on measuring the perioral region. Objectives: This study aimed to provide a standardized 3D anthropometric protocol for the perioral region. Methods: 38 female and 12 male Asians were recruited (mean age 31.6 ± 9.6 years). Two sets of 3D images using the VECTRA 3D imaging system were acquired for each subject, and two measurement sessions for each image were performed independently by two raters. 25 landmarks were identified, and 28 linear, 2 curvilinear, 9 angular and 4 areal measurements were evaluated for intrarater, interrater, and intramethod reliability. Results: Our results showed high reliability of 3D imaging-based perioral anthropometry by mean absolute difference (0.57 and 0.57 unit), technical error measurement (0.51 and 0.55 unit), relative error of measurement (2.18% and 2.44%), relative technical error of measurement (2.02% and 2.34%), and intraclass correlation coefficient (0.98 and 0.98) for intrarater 1 and intrarater 2 reliability; respectively 0.78 unit, 0.74 unit, 3.26%, 3.06% and 0.97 for interrater reliability; and respectively 1.01 unit, 0.97 unit, 4.74%, 4.57% and 0.95 for intramethod reliability. Conclusions: This standardized protocol utilizing 3D surface imaging technologies are feasible and highly reliable in perioral assessment. It could be further applied for diagnostic purpose, surgical planning and therapeutic effect evaluation in clinical practice in relation to perioral morphologies. Level of Evidence IV: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266. [ABSTRACT FROM AUTHOR]

Background: Single eyelids are common among Asians. It is not uncommon to see people with single eyelids raise their eyebrows to wide open their eyes. This results in frequent compensatory contraction of the frontalis muscle and thus leads to deep forehead wrinkles. Double-eyelid blepharoplasty creates a larger visual field. In theory, patients who receive the surgery will stop overusing the frontalis muscle. Therefore, the forehead wrinkles can be improved. Methods: 35 patients who underwent double-eyelid blepharoplasty were enrolled. FACE-Q forehead wrinkle assessment scale was adopted to evaluate the forehead wrinkles preoperatively and postoperatively. In addition, anthropometric measurements were taken to indirectly evaluate frontalis muscle contraction in maximum eye-opening position. Results: According to the FACE-Q scale, forehead wrinkles were improved after double-eyelid blepharoplasty, and the improvement was long-lasting in the 3-month follow-up. This was because the frontalis muscle contraction reduced after the surgery, as shown by the anthropometric measurements. Conclusion: This study used subjective and objective methods to prove that double-eyelid surgery improves forehead wrinkles. Level of Evidence IV: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266. [ABSTRACT FROM AUTHOR]

Background: Patients nowadays seek medical reduction of the upper trapezius muscle to achieve aesthetic pleasing necklines. Botulinum toxin type A (BoNT-A), a neurotoxin that reduces the force of muscle contraction, is widely used for shoulder contouring. However, detailed metrical data are lacking to guide clinical injection. Methods: Forty healthy young women were enrolled. All subjects were seated for point marking and measuring of trapezius muscle (TM) anatomical data. At marked locations, the thickness of the subcutaneous tissue, the thickness of the TM, the TM's anterior border, and the depth from the skin surface to the middle of TM were measured using a real-time ultrasound scanner. In addition, photos were taken to assess the shoulder area proportion and shoulder angle. Results: The shoulder area size among participants was 90 ± 15.2 cm2, the average area proportion was 0.55 ± 0.1, and the shoulder angle was 23.12° ± 2.9°. Ultrasonography data reveal that the thickness of the subcutaneous tissue, the thickness of the TM, and the depth from the skin surface to the middle of TM all thickened from the anterior line to the posterior line (P < 0.001). The length of the TM extending forward was 1.47 ± 0.4 cm (range 0.51–1.31 cm). To better evaluate the appearance of the trapezius muscle, we constructed a grading system correlated to shoulder angle. Conclusion: In this study, ultrasonography and photos were applied to assess the TM's anatomical features. Trapezius hypertrophy was described on a scale from 0 to 2 based on the shoulder angle. The findings of the present study provide a practical clinical guidance for precise and efficient BoNT-A administration. Level of Evidence IV: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266. [ABSTRACT FROM AUTHOR]

Background: Large calves are considered unattractive. Botulinum toxin type A (BoNT‐A), acting as a neurotoxin that reduces muscle contraction, is widely used in calf contouring. However, there lacks detailed anatomical data to guide clinical injection. Objectives: This study aims to use ultrasonography to provide depth information of the calf muscles and determine the morphological types of the gastrocnemius muscle. Methods: Ultrasound scanning was performed at six sites on 40 Chinese adults. A real‐time ultrasound scannerwas used to detect the depth of the gastrocnemius muscle and the soleus muscle. Results: The thickness of the gastrocnemius muscle and the soleus muscle was measured. The depth from the skin to the surface, the middle, and the bottom of the gastrocnemius muscle was measured. The morphology of gastrocnemius muscle was classified into four different types under ultrasound according to the muscle bulge pattern. Conclusion: This study took ultrasonic measurements of Asian calves to guide accurate and effective BoNT‐A injection in real clinical practice. [ABSTRACT FROM AUTHOR]

Background: The effect of radiotherapy (RT) for second primary malignancies (SPMs) among prostate cancer survivors is controversial. Methods: Applying logistic regression, competing risk analysis and propensity score matching method, this study analyzed clinical data from the Surveillance, Epidemiology, and End Results program to compare the risk for SPMs between patients receiving RT and non-RT. Results: In this study, prostate cancer patients treated with RT developed more SPMs in the anus, bladder, rectum, liver, lung and bronchus and lymphoma than non-RT groups. Conclusion: More intensive surveillance should be adopted for these cancers among prostate cancer survivors. [ABSTRACT FROM AUTHOR]

Background: Botulinum toxin (BoNT) injection is the most commonly performed procedure in cosmetic surgery. However, blind injection is unable to take individual anatomical variations into consideration, which is the main contributing factor to complications. Ultrasound (US) imaging was introduced to reduce complications and improve effects. This article will review uses of US in aesthetic BoNT injection. Method: A systematic electronic search was performed using the PubMed, MEDLINE, Web of science. Search terms were set to focus on aesthetic BoNT injection. Two independent reviewers subsequently reviewed the resultant articles based on strict inclusion and exclusion criteria. Selected manuscripts were analysed and grouped by procedure categories. Clinical cases were all performed by one plastic surgeon in our department. Results: The search finally retained 24 articles. Five procedural categories were identified, including masseter (n = 16), frontalis (n = 2), glabella complex (n = 2), trapezius (n=1), and gastrocnemius (n = 3). US imaging is practical and instructive for pre-operative assessments as in needle-type selection, injection point localization and depth setting, as well as post-operative follow-ups regarding injection feedback (for instance, the extent of muscle volume decreases). What's more, ultrasound-guided injection makes needle trajectory visualized so as for the needle to reach the target muscle in avoidance of potential damage to neurovascular bundle, gland or adjacent muscle. Conclusion: Muscles, such as masseter, frontalis, glabella complex, trapezius and gastrocnemius, and their adjacent structures can be well visualized using US, and as such, US can be a useful tool for a variety of pre-operative, intra-operative and post-operative procedures. Level of Evidence IV: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266. [ABSTRACT FROM AUTHOR]

Purpose: Keloids are caused by uncontrolled excessive proliferation of fibrous tissue. Multiple treatment strategies including steroid injection, surgical excision, laser therapy and radiation therapy have been reported. Few studies have evaluated the performance of plasma skin regeneration (PSR) in the treatment of keloid. This study aimed to evaluate the effectiveness of PSR combined with radiation therapy for keloids on different body parts. Patients and Methods: A total of 71 patients with 98 keloids were enrolled in this study. Keloids < 4 mm thick underwent single-dose PSR, while keloids ≥ 4 mm thick were administered compound betamethasone injection beforehand. Radiation therapy was administered after 24 hours and again 7 days later after PSR. The outcome was evaluated using the patient and observer scar assessment scale at 12 months post-treatment. Results: Patient-reported average scores for all keloids significantly decreased from 35.05± 9.94 to 21.84± 7.04 (p < 0.05). Keloids on face and neck, chest, and back responded better than those on shoulders and limbs. The recurrence rate was observed to be 15.3% (15 out of 98). Adverse effects were mild. Conclusion: PSR combined with radiation therapy is an effective and safe strategy to treat keloids. Location could be a factor that affects curative effects. [ABSTRACT FROM AUTHOR]

Background: The lip is of important aesthetic value and highly subjected to aging. Collecting anthropometric baseline data and understanding age‐related changes of labial morphology can help with diagnosis of deformity, assessment of aging, and planning of cosmetic procedures. Many studies have focused on Caucasians, while there is a lack of anthropometric data on Chinese women. Methods: A total of 169 women were enrolled in this cross‐sectional study and divided into four consecutive age groups. Linear distances, angles, and surface area data were obtained via stereophotogrammetry. Intergroup comparisons between different age groups were performed to find age‐related differences. Results: Lip width significantly increased with age while philtrum width seemed to show no obvious change. Cutaneous upper and lower lip height increased, lengthening the lip in the vertical dimension. Decrease of upper vermilion height and changes in angles indicated that aging process shortened the upper vermilion and flattened the vermilion border. Surface area also showed age‐related changes. Intergroup comparison showed no statistical significance in most variables between 20s and 30s or 30s and 40s, while age‐related changes in some variables were significant between 40s and 50s. Conclusion: This study provided anthropometric data of labial morphology across a wide age range. Aging process affected a variety of labial anthropometric variables. Age‐related changes accelerated after 40 among Chinese women. [ABSTRACT FROM AUTHOR]

Background: Full lips are beauty standards. Botulinum toxin A (BTA) paralyzes the orbicularis oris muscle to achieve the eversion of the lip, and thus makes the lip look plump. Aims: This study presents three cases of BTA injection to the lip and evaluates the possible changes in the labial morphology and the lip surface area. Methods: Three patients received a total of 4U BTA injection at the vermilion border of the upper lip. Vectra® H1 3D imaging system was used to capture 3D photographs of the lips before injection and two weeks after injection. Eight linear distances and the upper lip surface area were measured. Anthropometric measurements before and after injection were compared. Results: Patients displayed a larger upper vermillion height (P =.038) and a smaller cutaneous upper lip height (P =.024). There was a trend for a larger upper lip surface area, but not statistically significant (P =.109). Symptoms of slight perioral muscular palsy and mouth incompetence lasted about one month in three patients. Conclusion: BTA helps to enlarge the upper lip and shorten the philtrum. The BTA injection can be an option for lip enhancement with caution. [ABSTRACT FROM AUTHOR]

Background: Central lymph node metastasis (CLNM) occurs frequently in patients with papillary thyroid cancer (PTC), but performing prophylactic central lymph node dissection is still controversial. There are no reliable models for predicting CLNM. This study aimed to develop predictive models for CLNM by machine learning (ML) algorithms. Methods: Patients with PTC who underwent initial thyroid resection at our hospital between January 2018 and December 2019 were enrolled. A total of 22 variables, including clinical characteristics and ultrasonography (US) features, were used for conventional univariate and multivariate analysis and to construct ML-based models. A 5-fold cross validation strategy was used for validation and a feature selection approach was applied to identify risk factors. Results: The areas under the receiver operating characteristic curve (AUC) of 7 models ranged from 0.680 to 0.731. All models performed significantly better than US (AUC=0.623) in predicting CLNM (P<0.05). In decision curve, most of the models also performed better than US. The gradient boosting decision tree model with 7 variables was identified as the best model because of its best performance in both ROC (AUC=0.731) and decision curves. Based on multivariate analysis and feature selection, young age, male sex, low serum thyroid peroxidase antibody and US features such as suspected lymph nodes, microcalcification and tumor size > 1.1 cm were the most contributing predictors for CLNM. Conclusions: It is feasible to develop predictive models of CLNM in PTC patients by incorporating clinical characteristics and US features. The ML algorithm may be a useful tool for the prediction of lymph node metastasis in thyroid cancer. [ABSTRACT FROM AUTHOR]

Background: Patients who have previously undergone surgical resection of initial primary lung cancer (IPLC) are at high risk of developing second primary lung cancer (SPLC). There are still no standard treatments for SPLC. This study aimed to identify the prognostic factors and compare survival between the different SPLC treatment groups. Methods: SPLC patients in the Surveillance, Epidemiology, and End Results (SEER) database between 2007 and 2016 were retrospectively reviewed. Prognostic factors for SPLC were identified, using the least absolute shrinkage and selection operator (LASSO) regression and univariate Cox analysis to select variables for multivariate Cox analysis. Kaplan‐Meier method plus log‐rank test and restricted mean survival time (RMST) were used to compare survival outcome. Results: A total of 665 SPLC patients were finally enrolled into the study. Multivariate Cox regression analysis revealed that male vs. female (HR = 1.82, 95% CI: 1.29–2.59, P = 0.001), tumor size of SPLC ≥1 cm vs. <1 cm (HR = 1.80, 96% CI: 1.07–3.02, P = 1.028), IPLC characteristics of squamous cell carcinoma vs. adenocarcinoma (HR = 1.89, 95% CI: 1.17–3.04, P = 0.009), clinical stage II vs. stage I (HR = 2.60, 95% CI: 1.08–6.27, P = 0.033), and T2 stage vs. T1 stage (HR = 1.68, 95% CI: 1.04–2.72, P = 0.034) indicated worse survival. SPLC patients demonstrated a five‐year survival rate of 68.6% and a five‐year RMST of 49.4 months. The choice of surgical procedure (wedge resection, segmentectomy and lobectomy) for both IPLC and SPLC had no significant effect on prognosis (P > 0.05). Patients that received radiotherapy for SPLC also demonstrated similar survival when compared with those that underwent surgery (P > 0.05). Conclusions: Radiotherapy and sublobar resection can be considered reasonable alternative treatments for SPLC, especially when patients are unable to tolerate lobectomy. [ABSTRACT FROM AUTHOR]

Background: Lymph node metastasis (LNM) is difficult to precisely predict before surgery in patients with early-T-stage non-small cell lung cancer (NSCLC). This study aimed to develop machine learning (ML)-based predictive models for LNM. Methods: Clinical characteristics and imaging features were retrospectively collected from 1,102 NSCLC ≤ 2 cm patients. A total of 23 variables were included to develop predictive models for LNM by multiple ML algorithms. The models were evaluated by the receiver operating characteristic (ROC) curve for predictive performance and decision curve analysis (DCA) for clinical values. A feature selection approach was used to identify optimal predictive factors. Results: The areas under the ROC curve (AUCs) of the 8 models ranged from 0.784 to 0.899. Some ML-based models performed better than models using conventional statistical methods in both ROC curves and decision curves. The random forest classifier (RFC) model with 9 variables introduced was identified as the best predictive model. The feature selection indicated the top five predictors were tumor size, imaging density, carcinoembryonic antigen (CEA), maximal standardized uptake value (SUVmax), and age. Conclusions: By incorporating clinical characteristics and radiographical features, it is feasible to develop ML-based models for the preoperative prediction of LNM in early-T-stage NSCLC, and the RFC model performed best. [ABSTRACT FROM AUTHOR]