With the recent development of speech-enabled interactive systems using artificial agents, there has been substantial interest in the analysis and classification of voice disorders to provide more inclusive systems for people living with specific speech and language impairments. In this paper, a two-stage framework is proposed to perform an accurate classification of diverse voice pathologies. The first stage consists of speech enhancement processing based on the original premise, which considers impaired voice as a noisy signal. To put this hypothesis into practice, the noise lestral harmonic-to-noise ratio (CHNR). The second stage consists of a convolutional neural network with long short-term memory (CNN-LSTM) architecture designed to learn complex features from spectrograms of the first-stage enhanced signals. A new sinusoidal rectified unit (SinRU) is proposed to be used as an activation function by the CNN-LSTM network. The experiments are carried out by using two subsets of the Saarbruecken voice database (SVD) with different etiologies covering eight pathologies. The first subset contains voice recordings of patients with vocal cordectomy, psychogenic dysphonia, pachydermia laryngis and frontolateral partial laryngectomy, and the second subset contains voice recordings of patients with vocal fold polyp, chronic laryngitis, functional dysphonia, and vocal cord paresis. Dysarthria severity levels identification in Nemours and Torgo databases is also carried out. The experimental results showed that using the minimum mean square error (MMSE)-based signal enhancer prior to the CNN-LSTM network using SinRU, led to a significant improvement in the automatic classification of the investigated voice disorders and dysarhtria severity levels. These findings support the hypothesis that using an appropriate speech enhancement preprocessing has positive effects on the accuracy of the automatic classification of voice pathologies thanks to the reduction of the intrinsic noise induced by the voice impairment.