A nomogram for predicting sclerotherapy response for treatment of lymphatic malformations in children

Abstract

Purpose In this manuscript, we purposed to identify the prognostic factors for treatment of lymphatic malformations in children using polidocanol foam combined with pingyangmycin and to construct nomogram for predicting sclerotherapy response. Methods A retrospective analysis of 77 children having LMs who underwent sclerotherapy using polidocanol foam combined with pingyangmycin under ultrasound display from January 2017 to April 2020 was done. The clinical response was graded as excellent (≥ 90%), good (≥ 50%, < 90%), and poor (< 50%). More than 50% was considered as acceptable response. Prognostic factors were identified by Pearson’s Chi-square or Fisher’s exact test and multivariable logistic regression model was used to construct a nomogram to predict sclerotherapy response. The discrimination and calibration of nomogram were verified through the receiver operating characteristic cure and calibration plots. Results The mean number of treatment sessions was 3.1 (range, 1–6). Among 77 patients, 58 patients (75.3%) had excellent response to treatment (≥ 90%) and 68 patients (88.3%) had an acceptable response (≥ 50%, < 90%). Clinical disfigurement (P = 0.014), skin discoloration (P = 0.040), morphological subtype (P < 0.001) and extent of the lesion (P < 0.001) correlated with clinical response to sclerotherapy in LMs. Sclerotherapy response was predicted through nomogram constructed in this study, which shows good calibration and discrimination. Also, focal lesion and macrocystic or mixed morphological subtype lesion were seen more often in lower number of treatment sessions among the patients with excellent response. Conclusions An acceptable response to sclerotherapy using polidocanol foam combined with pingyangmycin was achieved in majority of LMs in children with extremely low complication rates. Nomogram based on the prognostic factors of sclerotherapy response for LMs in children was shown to possess an excellent performance to predict the probability of LMs sclerotherapy response.

Exploration and validation of key genes associated with early lymph node metastasis in thyroid carcinoma using weighted gene co-expression network analysis and machine learning

“…Additionally, we have established a web-based tool to access this nomogram's prediction model. The nomogram presented in this study can be easily utilized in clinical practice through our web-based tool, offering valuable resources and guidance for the formulation of clinical treatment and care strategies for THCA patients (87,88). Subsequently, employing an integrative analysis of three machine learning techniques, we identify ERBB3 as the key gene influencing LNM potential.…”