Validation of Rapid On-site Diagnostics in Early Stage Lung Cancer Sampling With a Higher Harmonic Generation Microscope

The gold standard for diagnosing all suspected malignancies, such as lung cancer, involves pathological examination using light microscopy, supplemented with immunohistochemical techniques and molecular (tissue) diagnostics. The official assessment of this tissue typically occurs after the procedure in which the tissue was obtained, through histological analysis (tissue analysis) or cytological analysis (cell analysis obtained through aspiration). Only then can certainty be provided regarding whether the obtained tissue is qualitatively adequate and/or quantitatively sufficient to make a diagnosis.

To ensure that complex and/or invasive procedures are performed with the correct outcome measures, intraoperative tissue diagnostics would be beneficial. In addition to postoperative assessment, additional intra-procedural evaluation of material for qualitative and quantitative estimation ('rapid onsite evaluation' (ROSE) and 'frozen section examination') is often used for complex and/or invasive procedures. Rapid on-site evaluation of cytopathology is always present during navigational bronchoscopy procedures, while frozen section examination is used for surgical tissue resection (such as in lung cancer surgery, as well as in oncological ENT and oesophageal surgery, among other organ and malignancy suspicions).

Navigational Bronchoscopy For patients with (a suspicion of) early-stage lung cancer, navigational bronchoscopy is a labour-intensive, complex procedure that has now become the new standard of care for diagnosing abnormalities suspicious for lung cancer. An endoscope and a subsequent navigable catheter are used to navigate through the natural airways to the suspicious lesion and obtain a biopsy to make a diagnosis. Since these are often peripheral small lesions, it is crucial to obtain tissue with millimetre precision. During navigational bronchoscopy, the previously mentioned ROSE technique is routinely used to receive feedback during the procedure on whether the obtained tissue is suitable for making a diagnosis. Histological and/or cytological tissue is collected using small instruments during navigational bronchoscopy, such as a slide filled with cell material. The cytology analyst assesses within minutes whether the material is representative by colouring and examining it under the microscope. Representativity means that the material contains enough for performing final pathology diagnostics to explain the abnormality. Since navigational bronchoscopy involves only small biopsies and tissue from the tumour itself or surrounding tissue is not obtained with certainty under direct visualization (biopsies are taken under the guidance of X-ray fluoroscopy), multiple and extensive biopsies are performed. In current clinical practice, routinely more than ten biopsies are taken per patient. In practice, we observe under intraoperative imaging that we could accurately reach the lesion in more than 90% of cases. In later follow-up, it turns out that in about 10-15% of procedures, we still cannot provide a correct diagnosis after analysing the obtained material.

Surgery Similar to ROSE during navigational bronchoscopy, frozen section examination is also requested during surgical procedures when there is no diagnosis available at the start of the procedure or when clinically relevant lymph node stations affect the intervention. Tissue is taken, and an employee brings it to the pathology department. The progress of the surgical procedure is paused. The tissue is frozen on the pathology department to cut sections, which are then coloured and assessed by the pathologist. Due to these activities, this means that the surgeon typically waits for more than half an hour for a pathology result during the surgery.

This approach is customary in numerous surgical procedures, not only in lung surgery but also in procedures for patients with oesophageal cancer, head and neck cancer, and other solid tumours elsewhere in the body.

In many clinical situations, faster and more detailed pathological feedback available during a procedure, providing immediate insights into the origin of abnormalities, could significantly improve clinical practice. Procedure wait times can be shortened, bridging the gap between accurately reaching a lesion and obtaining a diagnosis. It enables a more effective determination of the number of biopsies needed during diagnostic procedures such as navigational bronchoscopy and allows for more accurate intraoperative assessment of whether further therapy is necessary.

Higher Harmonic Generation Microscopy Higher Harmonic Generation Microscopy (HHGM) is a novel digital imaging technique that can generate (digital) microscopy images using laser light during a procedure. This technique produces non-invasive, label-free digital images of living tissue with sub-cellular resolution. The contrast in the images is caused by discontinuities, non-centrosymmetric molecular structures, or auto fluorescent organelles, generating higher optical harmonics and auto fluorescent photons, which are detected with a microscope. HHGM has been previously applied to image various tissue types, such as the breast, brain, and lungs. An earlier study demonstrated that HHGM provides high-quality images of lung tumours within minutes without fixation and staining. The resulting images revealed cell density, cellular and nuclear morphological features of the tumour, as well as extracellular matrix components (fibroblasts, collagen, and elastin), immune cells, red blood cells, and blood vessels. In a subsequent study at Amsterdam UMC, 109 biopsies from 47 patients were imaged within a median of 6 minutes (IQR = 3 minutes) after excision. The image quality was sufficient for diagnosing malignancy or non-malignancy in 97% of the biopsies, and 87% of the HHG images were correctly scored by pathologists.

Importantly, it was demonstrated that this optical imaging had no impact on the standard pathological processing and evaluation. The imaged tissue could be processed and assessed according to normal routine immediately after obtaining the images. In a previous study on human lung tissue at Amsterdam UMC (47 patients), histopathology images taken after imaging with the HHG microscope were evaluated by a pathologist. It was found that imaging with the HHG microscope also had no effect on the tissue or its accessibility. The laser power on the sample is 5 mW, equivalent to the power of a laser pointer.

The obtained tissues, before further processing in accordance with routine clinical processes, are placed in a special container for HHGM imaging. Both the small biopsies obtained, for example, during navigational bronchoscopy, and larger surgical resection tissues requiring longer scanning times are placed in these special containers, minimally moistened by adding a small amount of physiological saline (a drop to several drops) during imaging. This is done to improve imaging and prevent potential dehydration. This addition of physiological saline has no negative effect on the later assessment possibilities of the tissue by the pathologist.

The research protocol described here aims to collect a large number of HHGM images during routine procedures in daily clinical practice, creating a more extensive library of images for understanding HHGM images, training pathologists, and developing an AI algorithm for future computer-assisted diagnostics and/or automatic on-site diagnostics. The images collected during the procedure will not be used for clinical decision-making.