Noureddine Melikechi, dean of the Kennedy College of Sciences and professor at the University of Massachusetts Lowell, uses laser spectroscopy to advance biomedical research, focusing on the early detection of diseases such as epithelial ovarian cancer and Alzheimer’s. Spectroscopy spoke with Melikechi about how spectroscopic techniques are helping to tackle some of medicine’s most difficult diagnostic challenges.
What makes laser-induced breakdown spectroscopy (LIBS) a viable technique for detecting early biomarkers of epithelial ovarian cancer, and how does it compare to traditional optical methods?
LIBS is a versatile analytical technique that can analyze virtually any type of sample with minimal preparation. It provides simultaneous detection of multiple elements in a single test, offering a rapid and comprehensive elemental profile with no or minimal sample preparation. This makes LIBS highly efficient compared to many traditional methods.
LIBS, particularly Tag-LIBS, has emerged as a promising technique for the sensitive detection of early biomarkers for a variety of diseases, including epithelial ovarian cancer. By employing microparticles tagged with specific antibodies targeting cancer antigen 125 (CA125), Tag-LIBS enables detection of low concentrations of biomarkers in blood plasma. This approach has been shown to approach and potentially exceed the sensitivity enzyme-linked immunosorbent assay (ELISA).
More recently, scientists at the Brno University of Technology have further improved the sensitivity of Tag-LIBS using advanced nanoparticle-based labels and innovative pulse laser configurations. Collectively, these and other advances, demonstrate that LIBS can achieve excellent sensitivity and specificity. Tag-LIBS should be considered as a viable liquid biopsy technique for the detection and monitoring of ovarian cancer biomarkers, and possibly other pathologies.
Another avenue for the analysis of LIBS spectra is the integration of computational tools. LIBS generates high-dimensional data. Machine learning (ML) algorithms can identify subtle spectral features associated with ovarian cancer. This approach has enabled several research groups to distinguish between the elemental compositions of healthy and diseased samples in blood plasma. These innovations position LIBS as a good candidate that can contribute to advancements in liquid biopsy technologies. A key current limitation, however, lies in the often-restricted number of biomedical samples available to LIBS researchers, which can constrain statistical confidence and increase the risk of bias.
Your work spans both experimental and numerical methodologies—how are computational models shaping your approach to quantifying cancer biomarkers with laser-based techniques?
In our work, computational models are used as tools to extract meaningful, reproducible spectral features. We do this by searching for spectral signatures that may correlate with disease states.
What are the most promising optical or spectral signatures you've identified so far for early-stage cancer detection, and how are you validating them?
Our early work focused on cancer detection, but we have since expanded our efforts to Alzheimer’s disease, where accurate and early diagnosis remains a significant challenge. For both pathologies, we are developing liquid biopsy approaches based on the study of blood plasma. These not invasive and are well-suited for early-stage detection and longitudinal disease monitoring and, as a result, can offer substantial advantages over conventional tissue biopsies.
By leveraging a range of complementary spectroscopic techniques, we aim to uncover consistent signatures that effectively differentiate plasma samples from healthy individuals and those with disease. These include LIBS and inductively coupled plasma mass spectrometry (ICP-MS) to obtain the elemental composition of plasma, Fourier transform infrared (FT-IR) and/or Raman spectroscopy to gain molecular-level insights, and proteomics to explore changes in the protein landscape.
We recently demonstrated the potential of using elemental information in blood plasma to distinguish between individuals with Alzheimer’s disease and healthy controls. In another study, we have reported on proteins that can differentiate between these two groups of individuals. Our next step is to fuse this information to gain, hopefully, new insights about the disease. With the growing availability of powerful ML tools, fusing multiple data sets acquired with different techniques is becoming increasingly attractive. This integrated strategy not only strengthens confidence in the results and increases the likelihood of making new discoveries.
In your estimation, how close are LIBS diagnostics to receiving formal FDA approval for routine clinical use?
At present, LIBS-based diagnostics are still in the early stages of their translational journey. While recent advances in LIBS, such as Nanoparticle-enhanced LIBS (NELIBS), have demonstrated impressive sensitivity and specificity, several critical steps remain before LIBS can be considered for FDA approval and clinical adoption. These steps must include establishing standardized protocols for LIBS measurements in biomedical samples. This is non-trivial, given the variability in both biological matrices and laser-based instrumentation. Second, validating LIBs results through independent studies and clinical trials to test for reproducibility and clinical relevance. Only after these and other necessary steps are completed can LIBS be considered a viable liquid biopsy technique and a candidate for clinical application.
