A new rapid diagnostic system with ambient mass spectrometry and machine learning for colorectal liver metastasis

Recent Developments in Machine Learning for Mass Spectrometry

Statistical analysis and modeling of mass spectrometry (MS) data have a long and rich history with several modern MS-based applications using statistical and chemometric methods. Recently, machine learning (ML) has experienced a renaissance due to advents in computational hardware and the development of new algorithms for artificial neural networks (ANN) and deep learning architectures. Moreover, recent successes of new ANN and deep learning architectures in several areas of science, engineering, and society have further strengthened the ML field. Importantly, modern ML methods and architectures have enabled new approaches for tasks related to MS that are now widely adopted in several popular MS-based subdisciplines, such as mass spectrometry imaging and proteomics. Herein, we aim to provide an introductory summary of the practical aspects of ML methodology relevant to MS. Additionally, we seek to provide an up-to-date review of the most recent developments in ML integration with MS-based techniques while also providing critical insights into the future direction of the field.

Advances and Emerging Medical Applications of Direct Mass Spectrometry Technologies for Tissue Analysis

Offering superb speed, chemical specificity, and analytical sensitivity, direct mass spectrometry (MS) technologies are highly amenable for the molecular analysis of complex tissues to aid in disease characterization and help identify new diagnostic, prognostic, and predictive markers. By enabling detection of clinically actionable molecular profiles from tissues and cells, direct MS technologies have the potential to guide treatment decisions and transform sample analysis within clinical workflows. In this review, we highlight recent health-related developments and applications of direct MS technologies that exhibit tangible potential to accelerate clinical research and disease diagnosis, including oncological and neurodegenerative diseases and microbial infections. We focus primarily on applications that employ direct MS technologies for tissue analysis, including MS imaging technologies to map spatial distributions of molecules in situ as well as handheld devices for rapid in vivo and ex vivo tissue analysis.

Role of artificial intelligence in risk prediction, prognostication, and therapy response assessment in colorectal cancer: current state and future directions

Artificial Intelligence (AI) is a branch of computer science that utilizes optimization, probabilistic and statistical approaches to analyze and make predictions based on a vast amount of data. In recent years, AI has revolutionized the field of oncology and spearheaded novel approaches in the management of various cancers, including colorectal cancer (CRC). Notably, the applications of AI to diagnose, prognosticate, and predict response to therapy in CRC, is gaining traction and proving to be promising. There have also been several advancements in AI technologies to help predict metastases in CRC and in Computer-Aided Detection (CAD) Systems to improve miss rates for colorectal neoplasia. This article provides a comprehensive review of the role of AI in predicting risk, prognosis, and response to therapies among patients with CRC.

New Diagnostic Modality Combining Mass Spectrometry and Machine Learning for the Discrimination of Malignant Intraductal Papillary Mucinous Neoplasms

BACKGROUND

An intraductal papillary mucinous neoplasm (IPMN) is a pancreatic tumor with malignant potential. Although we anticipate a sensitive method to diagnose the malignant conversion of IPMN, an effective strategy has not yet been established. The combination of probe electrospray ionization-mass spectrometry (PESI-MS) and machine learning provides a promising solution for this purpose.

METHODS

We prospectively analyzed 42 serum samples obtained from IPMN patients who underwent pancreatic resection between 2020 and 2021. Based on the postoperative pathological diagnosis, patients were classified into two groups: IPMN-low grade dysplasia (n = 17) and advanced-IPMN (n = 25). Serum samples were analyzed by PESI-MS, and the obtained mass spectral data were converted into continuous variables. These variables were used to discriminate advanced-IPMN from IPMN-low grade dysplasia by partial least square regression or support vector machine analysis. The areas under receiver operating characteristics curves were obtained to visualize the difference between the two groups.

RESULTS

Partial least square regression successfully discriminated the two disease classes. From another standpoint, we selected 130 parameters from the entire dataset by PESI-MS, which were fed into the support vector machine. The diagnostic accuracy was 88.1%, and the area under the receiver operating characteristics curve was 0.924 by this method. Approximately 10 min were required to perform each method.

CONCLUSION

PESI-MS combined with machine learning is an easy-to-use tool with the advantage of rapid on-site analysis. Here, we show the great potential of our system to diagnose the malignant conversion of IPMN, which would be a promising diagnostic tool in clinical settings.

Ambient ms profiling of meningiomas: intraoperative oncometabolite-based monitoring

The primary method of initial treatment of meningiomas is radical neurosurgical intervention. Various methods of intraoperative diagnostics currently in development aim to improve resection efficiency; we focus on methods based on molecular profiling using ambient ionization mass spectrometry. Such methods have been proven effective on various tumors, but the specifics of the molecular structure and the mechanical properties of meningiomas raise the question of applicability of protocols developed for other conditions for this particular task. The study aimed to compare the potential clinical use of three methods of ambient ionization in meningioma sample analysis: spray from tissue, inline cartridge extraction, and touch spherical sampler probe spray. To this end, lipid and metabolic profiles of meningioma tissues removed in the course of planned neurosurgical intervention have been analyzed. It is shown that in clinical practice, the lipid components of the molecular profile are best analyzed using the inline cartridge extraction method, distinguished by its ease of implementation and highest informational value. Analysis of oncometabolites with low molecular mass is optimally performed with the touch spherical sampler probe spray method, which scores high in both sensitivity and mass-spectrometric complex productivity.

Probe electrospray ionization mass spectrometry-based rapid diagnosis of liver tumors

BACKGROUND AND AIM

Prompt differential diagnosis of liver tumors is clinically important and sometimes difficult. A new diagnostic device that combines probe electrospray ionization-mass spectrometry (PESI-MS) and machine learning may help provide the differential diagnosis of liver tumors.

METHODS

We evaluated the diagnostic accuracy of this new PESI-MS device using tissues obtained and stored from previous surgically resected specimens. The following cancer tissues (with collection dates): hepatocellular carcinoma (HCC, 2016-2019), intrahepatic cholangiocellular carcinoma (ICC, 2014-2019), and colorectal liver metastasis (CRLM, 2014-2019) from patients who underwent hepatic resection were considered for use in this study. Non-cancerous liver tissues (NL) taken from CRLM cases were also incorporated into the analysis. Each mass spectrum provided by PESI-MS was tested using support vector machine, a type of machine learning, to evaluate the discriminatory ability of the device.

RESULTS

In this study, we used samples from 91 of 139 patients with HCC, all 24 ICC samples, and 103 of 202 CRLM samples; 80 NL from CRLM cases were also used. Each mass spectrum was obtained by PESI-MS in a few minutes and was evaluated by machine learning. The sensitivity, specificity, and diagnostic accuracy of the PESI-MS device for discriminating HCC, ICC, and CRLM from among a mix of all three tumors and from NL were 98.9%, 98.1%, and 98.3%; 87.5%, 93.1%, and 92.6%; and 99.0%, 97.9%, and 98.3%, respectively.

CONCLUSION

This study demonstrated that PESI-MS and machine learning could discriminate liver tumors accurately and rapidly.

Fast Classification of Thyroid Nodules with Ultrasound Guided-Fine Needle Biopsy Samples and Machine Learning

A rapid classification method was developed for the malignant and benign thyroid nodules with ultrasound guided-fine needle aspiration biopsy (FNAB) samples. With probe electrospray ionization mass spectrometry, the mass-scan data of FNAB samples were used as datasets for machine learning. The patients were marked as malignant (98 patients), benign (110 patients) or undetermined (42 patients) by experienced doctors in terms of ultrasound, the B-Raf (BRAF) gene, and cytopathology inspections. Pairwise coupling was performed on 163 ions to generate 3630 ion ratios as new features for classifier training. With the new features, the performance of the multilayer perception (MLP) classifier is much better than that with the 163 ions as features directly. After training, the accuracy of the MLP classifier is as high as 92.0%. The accuracy of the single-blind test is 82.4%, which proved the good generalization ability of the MLP classifier. The overall concordance is 73.0% between prediction and six-month follow-up for patients in the undetermined group. Especially, the classifier showed high accuracy for the undetermined patients with suspicious for papillary carcinoma diagnosis (90.9%). In summary, the machine learning method based on FNAB samples has potential for real clinical applications.

Evaluation of CSTB and DMBT1 expression in saliva of gastric cancer patients and controls

Background

Gastric cancer (GC) is the fifth most common cancer and the third cause of cancer deaths globally, with late diagnosis, low survival rate, and poor prognosis. This case-control study aimed to evaluate the expression of cystatin B (CSTB) and deleted in malignant brain tumor 1 (DMBT1) in the saliva of GC patients with healthy individuals to construct diagnostic algorithms using statistical analysis and machine learning methods.

Methods

Demographic data, clinical characteristics, and food intake habits of the case and control group were gathered through a standard checklist. Unstimulated whole saliva samples were taken from 31 healthy individuals and 31 GC patients. Through ELISA test and statistical analysis, the expression of salivary CSTB and DMBT1 proteins was evaluated. To construct diagnostic algorithms, we used the machine learning method.

Results

The mean salivary expression of CSTB in GC patients was significantly lower (115.55 ± 7.06, p = 0.001), and the mean salivary expression of DMBT1 in GC patients was significantly higher (171.88 ± 39.67, p = 0.002) than the control.

Multiple linear regression analysis demonstrated that GC was significantly correlated with high levels of DMBT1 after controlling the effects of age of participants (R² = 0.20, p < 0.001).

Considering salivary CSTB greater than 119.06 ng/mL as an optimal cut-off value, the sensitivity and specificity of CSTB in the diagnosis of GC were 83.87 and 70.97%, respectively. The area under the ROC curve was calculated as 0.728. The optimal cut-off value of DMBT1 for differentiating GC patients from controls was greater than 146.33 ng/mL (sensitivity = 80.65% and specificity = 64.52%). The area under the ROC curve was up to 0.741.

As a result of the machine learning method, the area under the receiver-operating characteristic curve for the diagnostic ability of CSTB, DMBT1, demographic data, clinical characteristics, and food intake habits was 0.95. The machine learning model’s sensitivity, specificity, and accuracy were 100, 70.8, and 80.5%, respectively.

Conclusion

Salivary levels of DMBT1 and CSTB may be accurate in diagnosing GCs. Machine learning analyses using salivary biomarkers, demographic, clinical, and nutrition habits data simultaneously could provide affordability models with acceptable accuracy for differentiation of GC by a cost-effective and non-invasive method.

Applications of ambient ionization mass spectrometry in 2021: An annual review

Ambient ionization mass spectrometry (AIMS) has revolutionized the field of analytical chemistry, enabling the rapid, direct analysis of samples in their native state. Since the inception of AIMS almost 20 years ago, the analytical community has driven the further development of this suite of techniques, motivated by the plentiful advantages offered in addition to traditional mass spectrometry. Workflows can be simplified through the elimination of sample preparation, analysis times can be significantly reduced and analysis remote from the traditional laboratory space has become a real possibility. As such, the interest in AIMS has rapidly spread through analytical communities worldwide, and AIMS techniques are increasingly being integrated with standard laboratory operations. This annual review covers applications of AIMS techniques throughout 2021, with a specific focus on AIMS applications in a number of key fields of research including disease diagnostics, forensics and security, food safety testing and environmental sciences. While some new techniques are introduced, the focus in AIMS research is increasingly shifting from the development of novel techniques toward efforts to improve existing AIMS techniques, particularly in terms of reproducibility, quantification and ease-of-use.

Applications of Artificial Intelligence in Screening, Diagnosis, Treatment, and Prognosis of Colorectal Cancer

Colorectal cancer (CRC) is one of the most common cancers worldwide. Accurate early detection and diagnosis, comprehensive assessment of treatment response, and precise prediction of prognosis are essential to improve the patients’ survival rate. In recent years, due to the explosion of clinical and omics data, and groundbreaking research in machine learning, artificial intelligence (AI) has shown a great application potential in clinical field of CRC, providing new auxiliary approaches for clinicians to identify high-risk patients, select precise and personalized treatment plans, as well as to predict prognoses. This review comprehensively analyzes and summarizes the research progress and clinical application value of AI technologies in CRC screening, diagnosis, treatment, and prognosis, demonstrating the current status of the AI in the main clinical stages. The limitations, challenges, and future perspectives in the clinical implementation of AI are also discussed.

Artificial intelligence in the diagnosis and management of colorectal cancer liver metastases

Colorectal cancer (CRC) is the third most common malignancy worldwide, with approximately 50% of patients developing colorectal cancer liver metastasis (CRLM) during the follow-up period. Management of CRLM is best achieved via a multidisciplinary approach and the diagnostic and therapeutic decision-making process is complex. In order to optimize patients' survival and quality of life, there are several unsolved challenges which must be overcome. These primarily include a timely diagnosis and the identification of reliable prognostic factors. Furthermore, to allow optimal treatment options, a precision-medicine, personalized approach is required. The widespread digitalization of healthcare generates a vast amount of data and together with accessible high-performance computing, artificial intelligence (AI) technologies can be applied. By increasing diagnostic accuracy, reducing timings and costs, the application of AI could help mitigate the current shortcomings in CRLM management. In this review we explore the available evidence of the possible role of AI in all phases of the CRLM natural history. Radiomics analysis and convolutional neural networks (CNN) which combine computed tomography (CT) images with clinical data have been developed to predict CRLM development in CRC patients. AI models have also proven themselves to perform similarly or better than expert radiologists in detecting CRLM on CT and magnetic resonance scans or identifying them from the noninvasive analysis of patients' exhaled air. The application of AI and machine learning (ML) in diagnosing CRLM has also been extended to histopathological examination in order to rapidly and accurately identify CRLM tissue and its different histopathological growth patterns. ML and CNN have shown good accuracy in predicting response to chemotherapy, early local tumor progression after ablation treatment, and patient survival after surgical treatment or chemotherapy. Despite the initial enthusiasm and the accumulating evidence, AI technologies' role in healthcare and CRLM management is not yet fully established. Its limitations mainly concern safety and the lack of regulation and ethical considerations. AI is unlikely to fully replace any human role but could be actively integrated to facilitate physicians in their everyday practice. Moving towards a personalized and evidence-based patient approach and management, further larger, prospective and rigorous studies evaluating AI technologies in patients at risk or affected by CRLM are needed.

Utility of mass spectrometry and artificial intelligence for differentiating primary lung adenocarcinoma and colorectal metastatic pulmonary tumor

Background

Rapid intraoperative diagnosis for unconfirmed pulmonary tumor is extremely important for determining the optimal surgical procedure (lobectomy or sublobar resection). Attempts to diagnose malignant tumors using mass spectrometry (MS) have recently been described. This study evaluated the usefulness of MS and artificial intelligence (AI) for differentiating primary lung adenocarcinoma (PLAC) and colorectal metastatic pulmonary tumor.

Methods

Pulmonary samples from 40 patients who underwent pulmonary resection for PLAC (20 tumors, 20 normal lungs) or pulmonary metastases originating from colorectal metastatic pulmonary tumor (CRMPT) (20 tumors, 20 normal lungs) were collected and analyzed retrospectively by probe electrospray ionization‐MS. AI using random forest (RF) algorithms was employed to evaluate the accuracy of each combination.

Results

The accuracy of the machine learning algorithm applied using RF to distinguish malignant tumor (PLAC or CRMPT) from normal lung was 100%. The algorithms offered 97.2% accuracy in differentiating PLAC and CRMPT.

Conclusions

MS combined with an AI system demonstrated high accuracy not only for differentiating cancer from normal tissue, but also for differentiating between PLAC and CRMPT with a short working time. This method shows potential for application as a support tool facilitating rapid intraoperative diagnosis to determine the surgical procedure for pulmonary resection.

The phospholipid membrane compositions of bacterial cells, cancer cell lines and biological samples from cancer patients

While cancer now impacts the health and well-being of more of the human population than ever before, the exponential rise in antimicrobial resistant (AMR) bacterial infections means AMR is predicted to become one of the greatest future threats to human health. It is therefore vital that novel therapeutic strategies are developed that can be used in the treatment of both cancer and AMR infections. Whether the target of a therapeutic agent be inside the cell or in the cell membrane, it must either interact with or cross this phospholipid barrier to elicit the desired cellular effect. Here we summarise findings from published research into the phospholipid membrane composition of bacterial and cancer cell lines and biological samples from cancer patients. These data not only highlight key differences in the membrane composition of these biological samples, but also the methods used to elucidate and report the results of this analogous research between the microbial and cancer fields.

Lipid Profiling in Cancer Diagnosis with Hand-Held Ambient Mass Spectrometry Probes: Addressing the Late-Stage Performance Concerns

Untargeted lipid fingerprinting with hand-held ambient mass spectrometry (MS) probes without chromatographic separation has shown promise in the rapid characterization of cancers. As human cancers present significant molecular heterogeneities, careful molecular modeling and data validation strategies are required to minimize late-stage performance variations of these models across a large population. This review utilizes parallels from the pitfalls of conventional protein biomarkers in reaching bedside utility and provides recommendations for robust modeling as well as validation strategies that could enable the next logical steps in large scale assessment of the utility of ambient MS profiling for cancer diagnosis. Six recommendations are provided that range from careful initial determination of clinical added value to moving beyond just statistical associations to validate lipid involvements in disease processes mechanistically. Further guidelines for careful selection of suitable samples to capture expected and unexpected intragroup variance are provided and discussed in the context of demographic heterogeneities in the lipidome, further influenced by lifestyle factors, diet, and potential intersect with cancer lipid pathways probed in ambient mass spectrometry profiling studies.