SERS-Based Serum Molecular Spectral Screening for Benign and Malignant Pulmonary Proliferative Nodules

SERS-Based Serum Molecular Spectral Screening for Benign and Malignant Pulmonary Proliferative Nodules: A Multicenter, Open-Label, Double-Blind, Independent Data Analysis Clinical Trial

Pulmonary nodules are often an early indicator of lung cancer. With the widespread adoption of chest CT scans in routine physical examinations, an increasing number of pulmonary nodules are being detected, including a variety of small nodules such as inflammatory lesions, benign tumors, and malignant tumors. Currently, there is no unified international consensus on the diagnostic and treatment strategies for pulmonary nodules, as outlined by various global guidelines. Developing and implementing a comprehensive lung nodule and lung cancer screening program within public health management systems remains a complex and challenging endeavor. Advancing research and proposing lung cancer screening technologies that are highly sensitive, highly specific, simple, accessible, and cost-effective is an essential and pressing priority in modern healthcare.

Raman spectroscopy (RS), as a non-invasive and highly specific molecular detection technique, can be obtained at the molecular level to sensitively detect changes in biomolecules composed of proteins, nucleic acids, lipids, and sugars related to tumor metabolism in biological samples. The surface enhanced Raman spectroscopy (SERS) developed based on this technology is one of the feasible methods for high-sensitivity biomolecule analysis. Although SERS technology has shown good diagnostic efficacy in lots of preclinical studies in multiple tumors, it is limited to a generally small sample size and lacks external validation. There for, a clinical study of Raman spectra for tumor diagnosis is needed, which meets the following requirements: 1.An objective, fast and practical application of Raman spectral data processing is needed and deep learning method may be the best classification method; 2. It requires multicenter and large clinical samples to train deep learning diagnostic model, and verify its true efficacy through external data of prospective study.

In preliminary research, the investigators collected serum Raman spectroscopy data from a cohort of 191 patients with pulmonary nodules and developed an intelligent diagnosis system for distinguishing between benign and malignant pulmonary nodules using a machine learning model. The system achieved an accuracy of 89.7%. In order to obtain the highest level of clinical evidence and truly realize clinical transformation, this prospective, multi-center clinical study is designed to verify the intelligent diagnostic system for early diagnosis of prostate cancer.

In 2020, there were approximately 19.3 million new cancer cases and nearly 10 million cancer deaths worldwide. Among them, the number of new cases of lung cancer was approximately 2.21 million, accounting for 11.4% of the total number of new cancer cases. There were approximately 1.8 million deaths from lung cancer, ranking first among cancer deaths. In that year, the number of new cases of lung cancer in China was 816,000, accounting for 37% of the global total. In 2022, the latest number of lung cancer cases in China increased to 1.0606 million, and the number of lung cancer deaths was 733,000. In terms of the 5-year survival rate of lung cancer patients, the data released in 2018 in China (2012-2015) was 19.7%, which is still a large gap from the overall cancer 5-year survival rate of 46.6% by 2030 proposed in the "Healthy China Action-Cancer Prevention and Control Action Implementation Plan (2023-2030)". The prognosis of lung cancer at different stages is quite different. The 5-year survival rate of stage I lung cancer is 77%~92%, and the 5-year survival rate of stage IIIA~IVB lung cancer is 0~36%. Therefore, early diagnosis and treatment of lung cancer is the key to improving the 5-year survival rate of lung cancer and improving the prognosis of patients. However, most lung cancer patients are already in the late stage of lung cancer when they are diagnosed, and they have missed the opportunity for radical treatment. The main reason is that the primary and secondary prevention work is not done enough. It is necessary to develop advanced technologies and integrate them into the consensus guidelines for wide promotion.

Pulmonary nodules are early manifestations of lung cancer. With the popularization of chest CT screening in physical examination items, more and more lung nodules are found in physical examinations, including various types of small nodules, such as inflammatory lesions, benign tumor lesions, and malignant tumor lesions. In order to identify these types of nodules, clinicians often judge the two-dimensional imaging features of nodules based on their personal experience, such as plane diameter, whether there are burrs, lobes, calcification and other features to assess the probability of malignancy of lung nodules, but the accuracy of judging the benign and malignant nodules in this way is closely related to the experience and seniority of clinicians, and different doctors have different judgments on the same nodules. At present, there is no unified consensus on the diagnosis and treatment strategies of lung nodules recommended by multiple international consensus guidelines. In public health management facilities, the development and implementation of a comprehensive lung nodule lung cancer screening program is a complex and challenging task. Researching and proposing high-sensitivity and high-specificity, as well as simple, easy-to-popular and low-cost lung cancer screening technologies is an indispensable part of the healthcare system. In addition, due to the inconsistency of guidelines for the diagnosis and treatment strategies of lung nodules, the phenomenon of overdiagnosis and treatment of lung nodules is also common in clinical practice. How to avoid overdiagnosis and treatment needs more attention. Therefore, it is our responsibility to actively improve the accuracy of prediction of lung nodule canceration, reduce the rate of overdiagnosis and treatment, and increase the rate of early lung cancer intervention. Among the existing screening methods for early lung cancer, laboratory tests (especially the use of blood, urine or other liquid biopsies) are a low-cost, non-invasive and easily repeatable early prediction method compared with imaging or histopathological examinations, by detecting specific cancer biomarkers such as circulating tumor DNA, proteins, cancer metabolites, and even cell-derived exosomes and circulating tumor cells. However, there are still many challenges, including: 1) There are no effective and abundant tumor biomarkers for lung cancer; 2) There is no simple and feasible cancer detection method, especially in the asymptomatic stage; 3) There is no comprehensive analysis platform for large data sets to distinguish between healthy and lung cancer populations.

Raman spectroscopy (RS) is a non-invasive and highly specific material molecular detection technology that can be obtained at the molecular level to sensitively detect changes in biomolecules composed of proteins, nucleic acids, lipids and sugars related to tumor metabolism in biological samples. Surface-enhanced Raman spectroscopy (SERS) developed based on this technology is one of the feasible methods for highly sensitive biomolecular analysis technology. Although SERS technology has shown good diagnostic effects in a large number of preclinical studies of multiple tumors, it is limited by the generally small sample size and lack of external verification. Therefore, it is necessary to conduct clinical research on the use of Raman spectroscopy for tumor diagnosis, which meets the following requirements: 1. Objective, fast and practical Raman spectroscopy data processing methods are required, and machine and deep learning methods may be the best classification methods; 2. Multi-center, large-sample clinical samples are needed to train deep learning diagnostic models, and their true efficacy is verified by external data from prospective studies.

In previous study, the investigators collected serum Raman spectroscopy data from a cohort of 191 patients with pulmonary nodules, and built a Raman intelligent diagnosis system for benign and malignant pulmonary nodules based on a machine learning model. The accuracy of this intelligent diagnosis system reached 89.7%. In order to obtain the highest level of clinical evidence and truly achieve clinical transformation, this prospective, multi-center clinical study aims to verify the use of this intelligent diagnosis system for the early diagnosis of malignant pulmonary nodules.

Inclusion Criteria:

1. Participants with Lung cancer meeting the criteria of TNM (Ninth Edition);
2. Participants are willing to participate in this study and follow the research plan;
3. Participants or legally authorized representatives can give written informed consent approved by the Ethics Review Committee that manages the website.

Exclusion Criteria:

1. Participants with concomitant other malignant tumors;
2. Participants with missing baseline clinical data;
3. Participants with severe underlying lung diseases (such as bronchiectasis, bronchial asthma or COPD, etc.), or those with a history of occupational or environmental exposure to dust, mines or asbestos;
4. Participants who do not cooperate or refuse to participate in clinical trials at a later stage.