Elastographic Assessment of Xenograft Pancreatic Tumors

Abstract

High tissue pressures prevent chemotherapeutics from reaching the parenchyma of pancreatic ductal adenocarcinoma, which makes it difficult to treat this aggressive disease. Researchers currently use invasive probes to monitor the effectiveness of pressure-reducing therapies, but this practice introduces additional complications. Here, we hypothesize that Young's modulus is a good surrogate for tissue pressure because collagen density and hyaluoronic acid, the key features of the tumor microenvironment responsible for high tissue pressures, also affect modulus elastograms. To corroborate this hypothesis, we used model-based quasi-static elastography to assess how the Young's modulus of naturally occurring AsPc-1 pancreatic tumors varies with collagen density and hyaluoronic acid concentration. We observed that Young's moduli of orthotopically grown xenograft tumors were 6 kPa (p < 0.05) higher than that of their subcutaneously grown counterparts. We also observed a strong correlation between Young's modulus and regions within the tumors with high collagen (R² ≈ 0.8) and hyaluoronic acid (R² ≈ 0.6) densities. These preliminary results indicate that hyaluronic acid and collagen density, features of the pancreatic ductal adenocarcinoma tumor microenvironment responsible for high tissue pressure, influence Young's modulus.

Keywords: Model-based elastography; Pancreatic ductal adenocarcinoma; Total tissue pressure; Tumor microenvironment.

Fig. 1

(a) Photograph of experimental elastographic imaging system, consisting of a Sonixtouch tablet ultrasound scanner (BK Ultrasound, Peabody, MA, USA) equipped with a L40-8/12 probe (BK Ultrasound, Peabody) and a computer-controlled mechanical compression system. (b) Close-up of the transducer and excised xenograft tumor embedded in gelatin. (c) Rat with exposed orthotopically grown tumor.

Fig. 2

(a) Phantom containing cross-linked hyaluronic acid (HA) inclusion embedded in gelatin. We used the probe illustrated in (a) to measure interstitial fluid pressure in the HA inclusion. (b) Schematics of the phantom, ultrasound transducer and hole used to increase permeability. (c) Modulus elastograms obtained from phantoms that we submerged in phosphate-buffered saline for 0 and 72 h. The Young’s moduli of the gel surrounding the phantoms were 11.65 and 24.45 kPa. (d) Young’s modulus versus pressure within the HA inclusion that we embedded in a gel with Young’s modulus of 11.65 kPa. (e) Bar plot of Young’s moduli of the HA gels for the 11.65-, 18.92- and 24.45-kPa phantoms as a function of time.

Fig. 3

Images obtained from two orthotopically grownAsPc-1 tumors. (a, b) Ultrasound B-mode images. (c, d)Young’s modulus elastograms. (e, f) Masson’s trichrome-stained histologic images, (g, h) Hyaluronic acid-stained images.

Fig. 4

Images obtained from two subcutaneously grownAsPc-1 tumors. (a, b) Ultrasound B-mode images. (c, d)Young’s modulus elastograms. (e, f) Masson’s trichrome-stained histologic images. (g, h) Hyaluronic acid histologic images.

Fig. 5

Histograms of collagen density (a, b) and hyaluronic acid (HA) density (c, d) of subcutaneously (a, c) and orthotopically (b, d) grown tumors. (e) Boxplot of mean Young’s modulus (in kPa) of subcutaneous and orthotopic tumors.

Fig. 6

Co-registered histological images and modulus elastogram obtained from a subcutaneous tumor. The dotted red lines in (a) to (e) indicate the boundary of ultrasound scans. (a) Masson’s trichrome-stained histological image. (b) Hyaluronic acid (HA)-stained histological image. (c) Collagen density map. (d) HA density map. (e) Composite image of collagen and HA density maps. (f) Overlay of modulus elastogram and sonogram; red lines demarcate the tumor edge.

Fig. 7

Quantitative analysis of collagen distribution, hyaluronic acid distribution and Young’s modulus distribution of AsPc-1 tumors. (a) Collagen density versus Young’s modulus for rat (blue) and mouse tumors (black) obtained from regions of interests (ROIs) where collagen density were high, medium and low. (b) Young’s modulus versus hyaluronic acid (HA) density obtained from the same ROIs as in (a). (c) Young’s modulus versus HA density corresponding to regions of high, medium and low HA densities. (d) Young’s modulus as a function of collagen density using the same ROIs as in (c).

Fig. 8

(a) Total tissue pressure within an orthotopic tumor measured at regions with different collagen (blue stains) distributions. (b) Scatterplot of tissue pressure versus collagen density within AsPc-1 tumors (n = 24). We calculated the mean pressure from 10 statistically independent points within each tumor.