Spiking Neural Networks shade their light on the top of deep neural networks due to their sparse spikes derived energy efficiency. Though tradeoffs among performance accuracy and energy efficiency of deep SNNs to real world datasets, are bottleneck. Deep neural networks have succeeded decades with visual cognition towards cancer diagnose. Cervical cancer takes a significant statistics on teenage and adult women death. An already implemented medical image preprocessing and segmentation pipelines are commonly standalone software, optimized on a specific public data set. MRI of pelvis is superior to clinical cervical cancer staging. Nevertheless, heterogeneous tumor uptakes of imaging modality have ambiguous boundary with a limited contrast between targeting organs and the neighboring tissues. Therefore, in this study, we used deep spiking neural networks model towards cervical cancer stages identification, from invasive cc confirmed and stage reported MRI images dataset, which are acquired from Saint Paul Millennium College and hospital. Involving a radiologist, collected cervical MRI images are anonimized, enhanced, segmented and annotated via a sequential pipeline of medical images preprocessing and segmentation techniques. The proposed DSNN is pre trained with convolutional neural network for identifying lesion features then converted in to approximate SNN while varied regularization techniques are used to optimize performance among accuracy versus energy tradeoffs. Classification using spiking softmax is done. Also, we demonstrate the model on neuromorphic Nengo DL frame work. Accordingly, the proposed DSNN model performed 90% cervical cancer lesions identification accuracy and 84% for classifying cervical cancer stages. Relative to the state for the art which is an accuracy of 91.35% on CIFAR10, the homogenous nature of annotated dataset after augmentation and probabilistic nature of soft max on an inclusive nature of classes for cc stages affects the performance. Therefore, we recommend further studies to extend DSNNs towards large numbers of annotated MRI images for heterogeneous cancer types, substages, ages and concomitant pathology along with classification using non-probabilistic approach.