Diffusion-weighted imaging (DWI) is a functional MRI technique based on the Brownian motion of water molecules [[1], [2], [3]]. DWI plays a key role in hematologic and oncologic imaging, such as lesion detection, characterization, and response assessment via voxel-based quantification of the change in apparent diffusion coefficient (ADC) [[4], [5], [6]]. Regarding spinal imaging, DWI is the core sequence for the evaluation of metastatic bone and hematologic marrow diseases such as multiple myeloma (MM) [[7], [8], [9], [10]]. The single-shot echo-planar imaging (SS-EPI) sequence has been traditionally the most common choice for DWI of the spine in clinical practice because the acquisition of the entire k-space data within a single repetition time enables fast image acquisition and motion insensitivity [11,12]. These strong advantages are especially crucial for whole-spine and whole-body imaging, which require extensive craniocaudal coverage and, consequently, a long scan time. However, there are several major drawbacks to the SS-EPI sequence, namely, limited spatial resolution, susceptibility to magnetic field inhomogeneity, eddy current-induced geometric distortion, and Nyquist ghosting artifacts [[13], [14], [15], [16]]. In addition to these technical obstacles, the requirement for uniform fat suppression across a large field of view presents another challenge in SS-EPI-based diffusion imaging of the whole spine [13]. Magnetically inhomogeneous regions, such as the cervicothoracic junction and air-bone interface of the lung and thoracic spine, are prone to fail local fat suppression [17]. Bulky respiratory motions of the chest and abdomen that produce substantial B0 field distortion can degrade image quality despite the inherent motion insensitivity of SS-EPI [18]. Moreover, a sufficient signal-to-noise ratio (SNR) must be achieved for the delineation of a small osseous lesion and relevant anatomic structures of the spine and spinal cord. According to several previous studies, deep learning (DL) reconstruction has been increasingly implemented to improve the image quality of DWI of the liver, breast, and prostate gland [[19], [20], [21], [22], [23], [24]]. To our knowledge, however, no study has attempted to apply DL-reconstructed DWI (DL DWI) in the field of musculoskeletal and whole-spine imaging. Therefore, the purpose of this study was to investigate the feasibility of DL reconstruction and super resolution processing for improving the image quality of whole-spine DWI by conducting intraindividual comparisons of the qualitative and quantitative imaging parameters in DL DWI and conventional (CONV) DWI in patients with hematologic and oncologic diseases.