Needle biopsy is still "invasive" and it introduces the possibility of a sampling error if the needle is not placed in the area of a sample conincident with the pathology in question.  It isn't 100% effective since needle placement is critical. There is always a risk associated with introducing a needle into the body, even though that risk may be small.  Needles can damage unintended targets significantly.  Needle biopsies can come back being read quite accurately as normal when, in effect, the biopsied tissue was adjacent but not incorporating the abnormal tissue.  In that situation the pathologist correctly reads the sample as normal because that is the type of tissue that was submitted, but the abnormal tissue may have been only a millimeter to the left or right of the diseased tissue. The possibility of sampling error is the reason that needle biopsies are conducted by sampling representative multiple areas of the suspect site, but even though that increases the liklihood of retrieving a sample in the intended suspect zone, it isn't 100% a sure thing.

On the other hand, the Optical Coherence Tomography methodology appears to have its own limitations.  It would definitely be "invasive" unless the lesion happened to be on the skin surface since endoscopic methodolgy would have to be incorporated to do the optical tomogram.  That endoscope has to be introduced somewhere inside the body, so "invasive" depends on how you define it, but endoscopy is not risk free as someone might assume, although the risk is certainly reduced as opposed to taking an actual slice of tissue. Another limitation, as outlined in the article, is that the depth of this methodology is limited to the 2 to 3 mm range, at least as described in that 2020 article.