Pre-treatment growth and IGF-I deficiency as main predictors of response to growth hormone therapy in neural models

Abstract

Mathematical models have been applied in prediction of growth hormone treatment effectiveness in children since the end of 1990s. Usually they were multiple linear regression models; however, there are also examples derived by empirical non-linear methods. Proposed solution consists in application of machine learning technique – artificial neural networks – to analyse this problem. This new methodology, contrary to previous ones, allows detection of both linear and non-linear dependencies without assuming their character a priori. The aims of this work included: development of models predicting separately growth during 1st year of treatment and final height as well as identification of important predictors and in-depth analysis of their influence on treatment’s effectiveness. The models were derived on the basis of clinical data of 272 patients treated for at least 1 year, 133 of whom have already attained final height. Starting from models containing 17 and 20 potential predictors, respectively for 1st year and final height model, we were able to reduce their number to 9 and 10. Basing on the final models, IGF-I concentration and earlier growth were indicated as belonging to most important predictors of response to GH therapy, while results of GH secretion tests were automatically excluded as insignificant. Moreover, majority of the dependencies were observed to be non-linear, thus using neural networks seems to be reasonable approach despite it being more complex than previously applied methods.

Significance of Direct Confirmation of Growth Hormone Insensitivity for the Diagnosis of Primary IGF-I Deficiency

“…The results of present study, obtained in a specific group of children with severe deficit of height and severe IGFD, correspond to that obtained in our neural model of rhGH therapy effectiveness for children with short stature and wide spectrum of GH and IGF-I secretion, in which neither GH peak after pharmacological stimulation nor birth weight presented to be predictors of the attained FH [37].…”

A systematic review on machine learning in sellar region diseases: quality and reporting items

“…Three studies used the same image database such that only the latest published study was included. At last, this systematic review included 16 studies (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21) (Table 2), with the diagnosis of general pituitary neoplasms, acromegaly, Cushing’s disease, craniopharyngioma and growth hormone deficiency. More than half of the studies were published in the recent 2 years.…”

“…Sample size in these studies varied from tens to thousands. The majority of the studies (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 20) (76.5%) used the diagnosis of a specific disease as the outcome, only four studies (17, 18, 19, 21) tested on the treatment outcome. In the diagnostic studies, three studies (7, 12, 14) used image features to categorize magnetic resonance images (MRIs), two (6, 13) used face photos to predict acromegaly, two (15, 20) predicted growth home deficiency using serum proteins, two (10, 11) used histological spectrum to predict histology diagnosis, one (9) used serum proteins to predict pituitary adenoma and one (8) predicted surgical phase using videos.…”

“…In the diagnostic studies, three studies (7, 12, 14) used image features to categorize magnetic resonance images (MRIs), two (6, 13) used face photos to predict acromegaly, two (15, 20) predicted growth home deficiency using serum proteins, two (10, 11) used histological spectrum to predict histology diagnosis, one (9) used serum proteins to predict pituitary adenoma and one (8) predicted surgical phase using videos. In studies on treatment outcomes, one study (17) predicted poor early postoperative outcome, one (18) predicted gross-total resection, one (19) predicted response to somatostatin analogs and one (21) predicted growth after growth hormone treatment. All the outcomes were either dichotomized or categorical outcomes except one in the continuous form (21).…”

Verification of Road Accident Simulation Created With the Use of Pc-Crash Software

“…The literature presents many approaches to solving problems in the context of analysing images with AI [14][15][16][17]. The authors of AI-related papers have paid particular attention to the analysis of the sensitivity of relevant models [18,19].…”