Characterizing intestinal inflammation and fibrosis in Crohn's disease by photoacoustic imaging: feasibility study

Abstract

The pathology of Crohn's disease (CD) is characterized by obstructing intestinal strictures because of inflammation (with high levels of hemoglobin), fibrosis (high levels of collagen), or a combination of both. The accurate characterization of the strictures is critical for the management of CD. This study examines the feasibility of characterizing intestinal strictures by Photoacoustic imaging (PAI) without extrapolation from superficial biopsies. Ex vivo normal rat colon tissue, inflammatory and fibrotic intestinal strictures in rat trinitrobenzene sulfonic acid (TNBS) model were first differentiated by a PA-US parallel imaging system. Surgically removed human intestinal stricture specimens were afterwards imaged by a multiwavelength acoustic resolution PA microscope (ARPAM). The experiment results suggest that PAI is a potential tool for the diagnosis of the diseased conditions in intestinal strictures.

Keywords: (170.3880) Medical and biological imaging; (170.5120) Photoacoustic imaging; (170.7170) Ultrasound.

Fig. 1

Relative optical absorption spectra of major chemical components in biological tissue [12, 13]. The vertical dashed lines mark the relatively higher optical absorption of hemoglobin, lipid and collagen at 532, 1220 and around 1310 nm, respectively. a.u. = arbitrary units. Hb. = Hemoglobin.

Fig. 2

Experiment setup for rat colon tissue imaging. The sample tissue specimen was flattened and laid at the bottom of a sample holder submerged in water. Laser was delivered from the opening of the sample holder. PA waves were coupled into the wall of the sample tube and propagate into water. An US probe acquires the PA signals. A PA-US imaging system acquires both the PA and US images of the samples.

Fig. 3

Schematics of a prototype PAI system. Lens 1: plano-convex lens with $f=6 cm$. Lens 2: plano-convex lens with $f=1 cm$. Tran.: transducer. Amp.: amplifier. P/R: pulser/receiver. Sync.: synchronization. Blue ripples represent the PA signals. Red pattern above the sample is the beam profile formed by a ring of fibers. The PC screen shows the PA image of the vasculatures in a mouse ear acquired during the initial test of the system.

Fig. 4

(A) Estimated beam profile on the surface of sample. (B) Monte Carlo simulated 2D light fluence distribution in the plane including the transducer focusing axis at wavelength of 532 nm (absorption coefficient ${\mu }_a=0.07 c{m}^{-1}$, reduced scattering coefficient ${{\mu }_s}^{\text{'}}=15.3 c{m}^{-1}$, scattering anisotropy $g=0.95$). (C) 2D light fluence distribution at wavelength of 1220 nm (${\mu }_a=1.29 c{m}^{-1}$,${{\mu }_s}^{\text{'}}=5.5 c{m}^{-1}$, $g=0.95$). (D) 2D light fluence distribution at wavelength of 1310 nm (${\mu }_a=1.09 c{m}^{-1}$,${{\mu }_s}^{\text{'}}=5.0 c{m}^{-1}$, $g=0.95$). (E) Light fluence distribution along the transducer focusing axis (red dashed lines in B-D). (F) Simulated acoustic intensity map of the focused transducer.

Fig. 5

PA and US measurement of bowel wall tissues from normal and acute TNBS rats at 532 nm. (A) Representative photographs of ex vivo colon from normal and acute TNBS rats. The distal colon for imaging is denoted by black boxes. (B) US and PA images of the distal colons in (A). (C) Statistics of the average pixel intensity of 3 normal control versus 3 acute inflammatory samples. The upper and lower edges of the boxes are the 25th and 75th percentiles, respectively.

Fig. 6

PA and US measurement of acute vs. chronic inflammatory intestinal strictures in TNBS rat at 1370 nm. (A) Photographs of ex vivo colon from representative acute and chronic inflammatory samples. The distal colons for imaging are denoted by black boxes. (B) US and PA images of the distal colons in (A). (C) Statistics of the average pixel intensity of 10 pure inflammation samples versus 9 mixed inflamed and fibrotic bowel wall tissue samples. The boxplot approach in Fig. 5(C) is used here again. The dashed lines extend to the most extreme data points and do not consider outliers. The outliers are plotted as “+”.

Fig. 7

Representative PA molecular component images (A)(B) and histology images (B)(D) of human intestinal strictures. MC: mucosal layer; MS: muscle layer. The specimens were scanned at 532, 1220 and 1310 nm. The relative concentrations of hemoglobin, collagen and lipid were inversely solved by Eq. (1) and encoded in (A) and (C) in red, green and blue, respectively. The color scales were normalized to the maximum concentration of hemoglobin. Both samples have collagen components, which is confirmed in (B)(D) with Masson’s trichrome staining. The inflammations in the samples were confirmed by H&E staining, as shown in the insertions in (B) and (D) in red boxes. The insertions were taken at 20x magnification. Spatial correlations between the histology and the PA images are marked by green arrows (high collagen concentration), red arrows and dashed contours (neovasculatrues containing hemoglobin), and blue arrows (lipid deposition caused by edema). The porcine gel for fixing the samples with very low collagen concentration was also stained in blue yet with uniform texture and marked by black arrows.