Raman Spectroscopy Discloses Altered Molecular Profile in Thyroid Adenomas

An eXplainable Artificial Intelligence analysis of Raman spectra for thyroid cancer diagnosis

Raman spectroscopy shows great potential as a diagnostic tool for thyroid cancer due to its ability to detect biochemical changes during cancer development. This technique is particularly valuable because it is non-invasive and label/dye-free. Compared to molecular tests, Raman spectroscopy analyses can more effectively discriminate malignant features, thus reducing unnecessary surgeries. However, one major hurdle to using Raman spectroscopy as a diagnostic tool is the identification of significant patterns and peaks. In this study, we propose a Machine Learning procedure to discriminate healthy/benign versus malignant nodules that produces interpretable results. We collect Raman spectra obtained from histological samples, select a set of peaks with a data-driven and label independent approach and train the algorithms with the relative prominence of the peaks in the selected set. The performance of the considered models, quantified by area under the Receiver Operating Characteristic curve, exceeds 0.9. To enhance the interpretability of the results, we employ eXplainable Artificial Intelligence and compute the contribution of each feature to the prediction of each sample.

Raman spectroscopy combined with machine learning algorithms for rapid detection Primary Sjögren's syndrome associated with interstitial lung disease

BACKGROUND

Interstitial lung disease (ILD) is a major complication of Primary Sjögren's syndrome (pSS) patients.It is one of the main factors leading to death. The aim of this study is to evaluate the value of serum Raman spectroscopy combined with machine learning algorithms in the discriminatory diagnosis of patients with Primary Sjögren's syndrome associated with interstitial lung disease (pSS-ILD).

METHODS

Raman spectroscopy was performed on the serum of 30 patients with pSS, 28 patients with pSS-ILD and 30 healthy controls (HC). First, the data were pre-processed using baseline correction, smoothing, outlier removal and normalization operations. Then principal component analysis (PCA) is used to reduce the dimension of data. Finally, support vector machine(SVM), k nearest neighbor (KNN) and random forest (RF) models are established for classification.

RESULTS

In this study, SVM, KNN and RF were used as classification models, where SVM chooses polynomial kernel function (poly). The average accuracy, sensitivity, and precision of the three models were obtained after dimensionality reduction. The Accuracy of SVM (poly) was 5.71% higher than KNN and 6.67% higher than RF; Sensitivity was 5.79% higher than KNN and 8.56% higher than RF; Precision was 6.19% higher than KNN and 7.45% higher than RF. It can be seen that the SVM (poly) had better discriminative effect. In summary, SVM (poly) had a fine classification effect, and the average accuracy, sensitivity and precision of this model reached 89.52%, 91.27% and 89.52%, respectively, with an AUC value of 0.921.

CONCLUSIONS

This study demonstrates that serum RS combined with machine learning algorithms is a valuable tool for diagnosing patients with pSS-ILD. It has promising applications.

Clinical Use of Raman Spectroscopy Improves Diagnostic Accuracy for Indeterminate Thyroid Nodules

BACKGROUND AND OBJECTIVE

Molecular analysis of thyroid fine-needle aspiration (FNA) specimens is believed to improve the management of indeterminate nodules. Raman spectroscopy (RS) can differentiate benign and malignant thyroid lesions in surgically removed tissues, generating distinctive structural profiles. Herein, the diagnostic performance of RS was tested on FNA biopsies of thyroid gland.

DESIGN

Prospective, blinded, and single-center study.

METHODS

We enrolled 123 patients with indeterminate or more ominous cytologic diagnoses (TIR3A-low-risk indeterminate lesion, TIR3B-high-risk indeterminate lesion, TIR4-suspicious of malignancy, TIR5-malignant). All subjects were surgical candidates (defined by international guidelines) and submitted to FNA procedures for RS analysis. We compared RS data, cytologic findings, and final histologic assessments (as reference standard) using various statistical techniques.

RESULTS

The distribution of our study population was as follows: TIR3A:37, TIR3B:32, TIR4:16, and TIR5:38. In 30.9% of patients, histologic diagnoses were benign. For predicting thyroid malignancy in FNA samples, the overall specificity of RS was 86.8%, with 86.5% specificity in indeterminate cytologic categories. In patients with high-risk ultrasound categories, the specificity of RS increased to 87.5% for TIR3A, reaching 100% for TIR3B. Benign histologic diagnoses accounted for 72.9% of patients classified as TIR3A and 31.3% of those classified as TIR3B. Based on positive RS testing, unnecessary surgery was reduced to 7.4% overall (TIR3A-33.3%, TIR3B-6.7%).

CONCLUSIONS

This premier use of RS for thyroid cytology confirms its role as a valuable diagnostic tool and a valid alternative to molecular studies, capable of improving the management of indeterminate nodules and reducing unnecessary surgery.

Unraveling the Effects of Carotenoids Accumulation in Human Papillary Thyroid Carcinoma

Among the thyroid cancers, papillary thyroid cancer (PTC) accounts for 90% of the cases. In addition to the necessity to identify new targets for PTC treatment, early diagnosis and management are highly demanded. Previous data indicated that the multivariate statistical analysis of the Raman spectra allows the discrimination of healthy tissues from PTC ones; this is characterized by bands typical of carotenoids. Here, we dissected the molecular effects of carotenoid accumulation in PTC patients by analyzing whether they were required to provide increased retinoic acid (RA) synthesis and signaling and/or to sustain antioxidant functions. HPLC analysis revealed the lack of a significant difference in the overall content of carotenoids. For this reason, we wondered whether the carotenoid accumulation in PTC patients could be related to vitamin A derivative retinoic acid (RA) biosynthesis and, consequently, the RA-related pathway activation. The transcriptomic analysis performed using a dedicated PCR array revealed a significant downregulation of RA-related pathways in PTCs, suggesting that the carotenoid accumulation in PTC could be related to a lower metabolic conversion into RA compared to that of healthy tissues. In addition, the gene expression profile of 474 PTC cases previously published in the framework of the Cancer Genome Atlas (TGCA) project was examined by hierarchical clustering and heatmap analyses. This metanalysis study indicated that the RA-related pathways resulted in being significantly downregulated in PTCs and being associated with the follicular variant of PTC (FV-PTC). To assess whether the possible fate of the carotenoids accumulated in PTCs is associated with the oxidative stress response, the expression of enzymes involved in ROS scavenging was checked. An increased oxidative stress status and a reduced antioxidant defense response were observed in PTCs compared to matched healthy thyroids; this was possibly associated with the prooxidant effects of high levels of carotenoids. Finally, the DepMap datasets were used to profile the levels of 225 metabolites in 12 thyroid cancer cell lines. The results obtained suggested that the high carotenoid content in PTCs correlates with tryptophan metabolism. This pilot provided novel possible markers and possible therapeutic targets for PTC diagnosis and therapy. For the future, a larger study including a higher number of PTC patients will be necessary to further validate the molecular data reported here.

Update on International Medical Taxonomies of Biomarkers and Their Applications in Management of Thyroid Cancers

Biomarkers (BMs) are medical signs which can be precisely measured and reproduced. Mainly, BMs provide information on the likely disease which can occur in an individual. On the other hand, BMs also signal disease recurrence in patients receiving therapy. The U.S. Food and Drug Administration coupled with the National Institutes of Health and the European Medicines Agency have proposed two distinct procedures to validate BMs. These agencies have elaborated two glossaries to describe the role of BMs. The aim of this study was to investigate medical taxonomies adopted by different governmental agencies for BM validation. Additional goals were to analyze efficiencies of the validated and candidate BMs for thyroid cancers (TCs). Currently, thyroglobulin is validated for monitoring TCs. Sorafenib-tosylate, Doxorubicin-hydrochloride, Vandetanib, Cabozantinib-s-malate, Dabrafenib-mesylate, Trametinib-dimethyl-sulfoxide, Lenvatinib-mesylate, Pralsetinib and Selpercatinib are validated for TC treatment. Among candidate BMs for TC diagnosis, there are molecular combinations including BRAF, RAS, RET/PTC and PAX8-PPARγ mutations. Noteworthy are BRAF and RET/PTC alterations already validated as targets of Dabrafenib-mesylate, Pralsetinib and Selpercatinib. Finally, cellular expressions of c-met in nodal TC metastases have diagnostic imaging applications. On the basis of this analysis, BM taxonomies should have common standards internationally recognized. BMs show different efficiencies depending on their diagnostic or therapeutic use.

Recent Advances in the Label-Free Characterization of Exosomes for Cancer Liquid Biopsy: From Scattering and Spectroscopy to Nanoindentation and Nanodevices

Exosomes (EXOs) are nano-sized vesicles secreted by most cell types. They are abundant in bio-fluids and harbor specific molecular constituents from their parental cells. Due to these characteristics, EXOs have a great potential in cancer diagnostics for liquid biopsy and personalized medicine. Despite this unique potential, EXOs are not yet widely applied in clinical settings, with two main factors hindering their translational process in diagnostics. Firstly, conventional extraction methods are time-consuming, require large sample volumes and expensive equipment, and often do not provide high-purity samples. Secondly, characterization methods have some limitations, because they are often qualitative, need extensive labeling or complex sampling procedures that can induce artifacts. In this context, novel label-free approaches are rapidly emerging, and are holding potential to revolutionize EXO diagnostics. These methods include the use of nanodevices for EXO purification, and vibrational spectroscopies, scattering, and nanoindentation for characterization. In this progress report, we summarize recent key advances in label-free techniques for EXO purification and characterization. We point out that these methods contribute to reducing costs and processing times, provide complementary information compared to the conventional characterization techniques, and enhance flexibility, thus favoring the discovery of novel and unexplored EXO-based biomarkers. In this process, the impact of nanotechnology is systematically highlighted, showing how the effectiveness of these techniques can be enhanced using nanomaterials, such as plasmonic nanoparticles and nanostructured surfaces, which enable the exploitation of advanced physical phenomena occurring at the nanoscale level.