Biomedical Engineering Reference
In-Depth Information
and skin diseases are apparent. We used partial least squares (PLS) regression of
the measured Raman spectra to derive the biochemical constituents in each lesion
and then used the linear discriminant analysis (LDA) to classify the skin diseases.
Our preliminary results showed that malignant melanoma can be differentiated from
other pigmented benign lesions with a diagnostic sensitivity of 91% and specificity
of 75%, while precancerous and cancerous lesions can be differentiated from benign
lesions with a sensitivity of 97% and a specificity of 78%, based on leave-one-out
cross-validation (LOOCV) analysis.
1.4.2
Lung Cancer Diagnosis
Normal and cancerous bronchial tissue have been studied in vitro using the real-
time skin Raman spectroscopic system [ 38 ]. Bronchial tissue specimens (12 normal,
10 squamous cell carcinoma (SCC), and 6 adenocarcinoma) were obtained from
10 patients with known or suspected malignancies of the lung. High-quality Raman
spectra from human bronchial tissues could be obtained within 5 s as shown in
Fig. 1.20 . Raman spectra differed significantly between normal and malignant tumor
tissue, with tumors showing higher percentage signals for nucleic acid, tryptophan,
and phenylalanine and lower percentage signals for phospholipids, proline, and
valine, compared to normal tissue. The ratio of Raman intensities at 1,445 to
1;655 cm 1 provided good differentiation between normal and malignant bronchial
tissue (p < 0:0001).
In vivo lung cancer diagnosis has been evaluated using the endoscopic real-time
Raman spectroscopic system (Fig. 1.14 )[ 39 , 40 ]. The physician identified suspect
lung sites of concern using both white light and fluorescence imaging. After these
sites were identified, the Raman probe was positioned with the aid of the lowered
excitation intensity (10%) beam spot, and then the Raman spectra were obtained.
Biopsies were then taken from the same sites and sent for histological evaluation.
Based on the pathologist's reports, the lesions were classified into normal, mild
dysplasia, moderate dysplasia, severe dysplasia, carcinoma in situ (CIS), or a
specific tumor type such as squamous cell carcinoma (SCC). In addition to the
normal from the pathology, Raman spectra were obtained from sites that appeared
completely normal in the physician's opinion, but no biopsies were obtained.
Contrary to the in vitro measurements [ 38 ], autofluorescence dominated the LF
range, which made it difficult to discern the Raman peaks. In the high-frequency
range, Raman peaks could be clearly seen from the raw spectra (Fig. 1.21 a).
This was the first in vivo Raman spectra of lung tissue. Pure Raman spectra
could be obtained after fluorescence removal for the whole range of measurement
(Fig. 1.21 b). The dominated Raman peaks were marked in Fig. 1.21 b. The broad
peaks near 1;658 cm 1 were due to a combination amide I (C D O) vibrations
and water molecule bending motions. The intense peaks at 2;900 cm 1 were
assigned to a combination of CH 2 antisymmetric (2;880 cm 1 )andCH 3 symmetric
(2;935 cm 1 ) stretching modes of phospholipids and proteins. The broad peaks near
3;300 cm 1
were most likely due to overlapping the symmetric and antisymmetric
Search WWH ::




Custom Search