Fig 1.
Histology photographs of retinoblastoma and uveal melanoma.
(A) Retinoblastoma. (B) uveal melanoma. The arrows in (A) mark the calcification spots forming the heterogeneous tissue architecture in retinoblastoma. The slides were prepared by H&E staining. Images were taken at 4x magnification.
Fig 2.
Optical absorption spectra of major molecular components in intraocular tumors [12, 34].
The black arrows indicate that the optical absorption spectrum of melanoma constantly decrease with respect to the increase of wavelengths. The optical absorption spectrum of the microcalcification shows a rapid decrease between the first two magenta arrows and a large curvature at the second magenta arrow. Both changes were captured in the PA images of retinoblastoma in the results section.
Fig 3.
Illustration of experiment systems.
(A) mouse experiment in vivo. (B) human eye globe imaging ex vivo.
Fig 4.
B-scan PA (acquired at 720 nm) and US images of retinoblastoma tumor in mice.
The yellow arrows mark the contour of the eye globe. The cyan arrows mark the skull of the mouse. The top of the PA images show bright artifacts due to the absorption of backscattered optical energy by the US probe surface. Scale bar: 1 cm.
Fig 5.
Relative optical absorption spectra complied by combining data acquired animal in vivo and human tissues ex vivo.
(A) retinoblastoma in mice in vivo. (B) Retinoblastoma in human ex vivo. (C) Uveal melanoma in human ex vivo.
Fig 6.
Representative US and PA images of ex vivo human ocular globes with tumors.
Images in the upper row are from a retinoblastoma tumor. Images in the lower row are from a uveal melanoma tumor. The yellow contours mark the tumors. The pixel-wise optical spectral parameters [a, b and c in Eq 1] and PASA slopes within the tumor regions were calculated and coregistered to the US images. The units of the color axes are arbitrary. Scale bar: 1 cm.
Table 1.
Statistics of the optical absorption spectral parameters derived with linear model, i.e. 1st order polynomial.
Table 2.
Statistics of the optical absorption spectral parameters derived with 2nd order polynomial.
Table 3.
Statistics of the optical absorption spectral parameters derived with 3rd order polynomial.
Fig 7.
Statistical study of the quantitative features.
(A)(B)(C) are the a, b and c values in Eq 2 derived from the total optical absorption spectra in Fig 5B and 5C.
Fig 8.
Illustration of the shadowing phenomenon in PAI.
(A) Transversal cross-section of the tumor in US image. (B) PA image shows a larger tumor profile [marked by magenta contour in (D)] than that in US image [cyan contour in (C)]. (C) Sagittal cross-section of the dome-shaped uveal melanoma in US. (D) Illustration of the imaging planes in (A), (B) and (C). Since US transmission and reception were both oriented at the cyan plane in (D), only the tumor profile within the plane was shown. The larger PA tumor profile was caused by: 1) PA signals generated within the whole tumor volume due to the penetrating illumination; and 2) PA measurements using only the receiving mode of the US array, which covered the 3D volume marked by the magenta contour in (D). The optical attenuation by the apex of the tumor cast a shadow at the tumor location, marked by the dashed circle in (B).
Fig 9.
PASA of the human intraocular tumors.
(A) Representative averaged PA signals spectra at 720 nm derived from the tumor samples. The solid lines are the linear fits to the signal power spectra within the probe frequency bandwidth (7-15MHz). Frequency range of 6.8–14.8 MHz was actually observed due to the limited sampling resolution. The PASA slope is tan(θ). Midband-fit is the magnitude of the linear model at the center of the observed frequency range (10.8 MHz for this case). (B) Statistics of the PASA slopes of retinoblastomas and uveal melanomas. The slopes derived from uveal melanomas have an average of -4.5 and standard deviation of 0.56. The slopes derived from retinoblastomas have an average of -5.6 and standard deviation of 0.60. Two-tailed t-test between the two groups has shown a p-value of 0.025.