Prediction of postpartum depression using multilayer perceptrons and pruning.

OBJECTIVE. The main goal of this paper is to obtain a classification model based on feed-forward multilayer perceptrons in order to improve postpartum depression prediction during the 32 weeks after childbirth with a high sensitivity and specificity and to develop a tool to be integrated in a decision support system for clinicians. MATERIALS AND METHODS. Multilayer perceptrons were trained on data from 1397 women who had just given birth, from seven Spanish general hospitals, including clinical, environmental and genetic variables. A prospective cohort study was made just after delivery, at 8 weeks and at 32 weeks after delivery. The models were evaluated with the geometric mean of accuracies using a hold-out strategy. RESULTS. Multilayer perceptrons showed good performance (high sensitivity and specificity) as predictive models for postpartum depression. CONCLUSIONS. The use of these models in a decision support system can be clinically evaluated in future work. The analysis of the models by pruning leads to a qualitative interpretation of the influence of each variable in the interest of clinical protocols.

Selective use of postoperative neck radiotherapy in oral cavity and oropharynx cancer: a prospective clinical study.

BACKGROUND In cervical postoperative radiotherapy, the target volume is usually the same as the extension of the previous dissection. We evaluated a protocol of selective irradiation according to the risk estimated for each dissected lymph node level. METHODS Eighty patients with oral/oropharyngeal cancer were included in this prospective clinical study between 2005 and 2008. Patients underwent surgery of the primary tumor and cervical dissection, with identification of positive nodal levels, followed by selective postoperative radiotherapy. Three types of selective nodal clinical target volume (CTV) were defined: CTV0, CTV1, and CTV2, with a subclinical disease risk of <10%, 10-25%, and 25% and a prescribed radiation dose of <35 Gy, 50 Gy, and 66-70 Gy, respectively. The localization of node failure was categorized as field, marginal, or outside the irradiated field. RESULTS A consistent pattern of cervical infiltration was observed in 97% of positive dissections. Lymph node failure occurred within a high-risk irradiated area (CTV1-CTV2) in 12 patients, marginal area (CTV1/CTVO) in 1 patient, and non-irradiated low-risk area (CTV0) in 2 patients. The volume of selective lymph node irradiation was below the standard radiation volume in 33 patients (mean of 118.6 cc per patient). This decrease in irradiated volume was associated with greater treatment compliance and reduced secondary toxicity. The three-year actuarial nodal control rate was 80%. CONCLUSION This selective postoperative neck irradiation protocol was associated with a similar failure pattern to that observed after standard neck irradiation and achieved a significant reduction in target volume and secondary toxicity.