Both overdiagnosis and underdiagnosis of melanoma cases lead to severe consequences. The early detection of melanoma is crucial as this cancer form can spread to other organs. On the other side, skin lesions which pose no danger to the patient can resemble melanoma features, leading to unnecessary costly biopsy procedures, triggering patient worries. Artificial Intelligence (AI) is a valuable tool for skin lesion classification, assisting dermatologists in their diagnosis. Recent studies explore the application of AI models using dermatoscopic or macroscopic image data. Image data comprises visual information from which AI models can extract essential features that may have been missed during the examination. However, multiple benign lesion types, like reed nevus or atypical nevus, can resemble melanoma cases, which can lead to misclassification. Millimetre waves are considered one of the most promising non-invasive techniques in skin cancer diagnosis. This technique can help image classifiers achieve higher accuracy levels by providing information about the cells. In this work, we investigate the potential of using multimodal data (image and millimetre wave data) to differentiate multiple skin lesion types. To perform the experiment, image and millimetre wave data from multiple lesions were collected at the Centre for Skin Diseases at University Hospital Bonn. The features of both modalities were extracted, after which Linear Discriminant Analysis was performed. The integrated multimodal data demonstrated better class differentiation than the single-modality approach.