A pilot trial of intravital microscopy in the study of the tumor vasculature of patients with peritoneal carcinomatosis

Abstract

Aberrancies in the tumor microvasculature limit the systemic delivery of anticancer agents, which impedes tumor response. Using human intravital microscopy (HIVM), we hypothesized that HIVM would be feasible in patients with peritoneal carcinomatosis (PC). During cytoreductive surgery with hyperthermic intraperitoneal chemotherapy for PC, HIVM was performed in both tumor and non-tumor areas. The primary outcome was HIVM feasibility to measure vessel characteristics. We secondarily evaluated associations between HIVM vessel characteristics and oncologic outcomes (RECIST response to neoadjuvant therapy and disease-specific survival). Thirty patients with PC were enrolled. Nineteen patients (63.3%) received neoadjuvant therapy. HIVM was feasible in all patients. Compared to non-tumor (control) areas, PC areas had a lower density of functional vessels, higher proportion of non-functional vessels, smaller lumenal diameters, and lower blood flow velocity. Qualitative differences in these vessel characteristics were observed among patients who had partial response, stable disease, or progressive disease after receiving neoadjuvant therapy. However, no statistically significant relationships were found between HIVM vessel characteristics and oncologic outcomes. These novel findings comprise the first-in-human, real-time evidence of the microscopic differences between normal and tumor-associated vessels and form the basis for our larger, ongoing clinical trial appropriately powered to determine the clinical utility of HIVM (NCT03823144).

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1

HIVM images of tumor and non-tumor vessels among individual. The right side panel shows examples of PC-associated tumor vessels from different histologies, including appendiceal adenocarcinoma (A), ovarian serous carcinoma (B), and mesothelioma (C). Yellow arrowheads highlight aberrantly arranged PC-associated vessel architecture (such as the acute hair-pin turn in part B), and yellow arrows highlight non-functional vessels as noted by the absence of fluorescein uptake. Non-tumor associated vessels from the same patients in the left side panel show normal, streamlined blood vessels, fewer non-functional vessels, and a higher number of functional (fluorescent) blood vessels. Red arrows highlight functional, normal blood vessels. Non-tumor areas were found to have some non-functional blood vessels, and conversely tumor areas were also found to have some functional blood vessels (red arrows).

Figure 2

HIVM observations from non-tumor (left panel) and tumor (right panel) areas obtained from a current smoker (A) and from a patient who received pre-operative radiotherapy (B). There was a similar proportion of non-functional vessels between the tumor and non-tumor areas in the current smoker, suggesting pre-existing background changes in the microvasculature due to smoking use. In patients who received pre-operative radiotherapy (n = 3), there was a high proportion of non-functional vessels observed (64.7%). These findings suggest that radiation-induced changes had affected the tumor microvasculature.

Figure 3

Examples of HIVM tumor observations for patients who had partial response (A), stable disease (B), and progressive disease (C) by standard RECIST criteria. The left side panels depict baseline imaging (CT scan or MRI) prior to neoadjuvant therapy. The middle panels depict follow-up imaging after neoadjuvant therapy just prior to undergoing CRS-HIPEC. Yellow arrows highlight the areas of disease. The right side panels depict HIVM PC-associated vessel observations. Patients who had achieved partial response had the highest density of functional vessels, lowest density of non-functional vessels, and the lowest proportion of non-function vessels when compared to patients who had stable disease or progressive disease. Although the subset analysis of patients who had neoadjuvant therapy did not show significant associations between RECIST response and tumor HIVM vessel characteristics, these individual patient observations provide early evidence that there may be a correlation between tumor response and tumor-associated vessels.