https://www.ngsci.org/updates/detecting-tuberculosis-using-ai

The current histology scanners on the market are great. They are the driving force in modern digital pathology. However, as a #cytologist, are you satisfied with the images from these scanners??

The histology scanners were designed to capture images from a smooth surface by a thin cut from the tumors or tissues. So it works like a Xerox machine making photocopies from a piece of paper. It looks great!!

What if you need digital images from a rough surface with all kinds of big and small players on it?? That is the case for #cytology. We have small players (mainly neutrophils and lymphocytes) and big players like mesothelial cells and #adenocarcinomacells in groups or single-cell patterns. So how could a 2-D imaging system accommodate different diameters of cells and produce fine cytology images for diagnosis and for #AI/ #ML training?? What is the focal point?

We answered the call. #HighMagfinicationCytology (#HMC) is an automatic microscopic imaging system designed for cytology and hematology. See our sample images in JPG format and forget about the digital storage fee for large histology slides.

Labeling was used extensively in building AI/ML algorithms. The idea is to apply the minds and knowledge of professionals to tell the AI/ML models to learn from the experts. Usually, in medical AI, the labeling identifies and marks the "Area of Interest" (AoI) for AI/ML as "positive". Army of pathologists have dedicated their time and efforts to this labor-intensive work, and we really appreciate it.

However, there is a problem: what if the professionals disagree with the EXACT AoIs?? We asked 4 cytologists, all with 10+ years of experience in clinical routine, to label cells to be likely "abnormal" (examples below). See the differences? How could an accurate AI/ML model be built with such inconsistency? Well, the current solution is to introduce MORE cytologists (as referees) to make judgments, but only to make the whole thing out of control.

The Solution: We are leading a team of #cytologists and #AI #ML computer experts to solve this problem using "#labelfree" #unsupervisedlearning. Would love to reveal our outcome in the near future. But I can share with you: the key to a successful label-free training is #HighMagnificationCytology (#HMC) images or otherwise garbage in garbage out #GIGO

A lot of innovation, new directions, and applications. 

Take home message : Accelerated Computing and Generative AI

The 13th Emirates Pathology, Digital Pathology & Cancer Conference from December 15-17, 2023, in Dubai, UAE & Virtual. Come join us for a stimulating discussion.
Note: Publish the full paper, articles, manuscripts, case study, refer to our supporting journal; Global Journal of Pathology & Laboratory Medicine [GJPLM]

WhatsApp us at https://lnkd.in/gnzBP_ns
Submit your Reasech paper: https://lnkd.in/g4w3XfY

Anything we can do today to improve the cytology screening accuracy after 13 years? It is definitely a "Yes". High Magnification Cytology (#HMC) that is better (4k image quality), faster (more FOVs to review), prettier (top color fidelity), and more accurate (latest #AI/ML technology) is the solution. Let's work together!!

Despite many advances in medicine, however, there are still many late-stage patients when diagnosed and have missed the best surgical treatment period. Late-stage patients are often in poor condition and cannot sustain invasive tests. Thus, cytology smears are of great benefit to patients to diagnose, classify, and stage tumors through minimally invasive approaches by examining pleural effusion and other fluids. The pathologists (or cytopathologists, to be exact) will manually observe these smears under microscopes to find cancer at cell levels.

Both sides aim to reduce the cancer mortality rate by half by 2050, in line with Biden's #cancermoonshot national policy. Taiwan will focus on breast cancer, lung cancer, liver cancer, etc.., and save 25,000 lives every year.

Early detection is always the key to reducing patient complications and mortality rates. We believe AI/ML should be implemented a lot faster than we have planned, and #HighMagnificationCytology can be a solution to make an early diagnosis of metastatic cancer; thus, patients receive appropriate #anticancertherapy to increase their survival rates.

Saving cancer patients one cell at a time EARLY!!! #scienceandtechnology #ai #ml #HMC #machinevision

https://lnkd.in/gK9JvA7g

Single-cell pattern adenocarcinoma cells, unlike cohesive morphology, are difficult to detect with low magnification screening under the microscope. These single cells hide among the background cells and are often confused with reactive mesothelial cells. A study (Guleria et al. 2021) showed, of 103 cases reported as either suspicious or positive for malignancy on PFC, 29 had a predominant single-cell pattern. Of these, 13 (44.8%) were primary lung carcinoma. The rest were metastasis from the ovary (5; 17.2%), breast (2; 6.9%), stomach (2; 6.9%), lymphoma (1; 3.5%), and Ewing's sarcoma (1; 3.5%). The take: Single-cell pattern of pulmonary adenocarcinoma is more common than popularly believed.

The solution: High Magnification Cytology (#HMC) with #AI and #MachineLearning is the only weapon to win the morphologic challenges of lung cancer diagnosis.

Traditional cytology screening for cancer is performed by pathologists by hand (image on the left). Difficult for humans or AI + ML to make a diagnosis because of poor image fidelity. Wellgen Medical's high magnification cytology (image on the right) is a game changer. High Magnification Cytology enables high-resolution machine learning for the early detection of cancers and bacteria at the cellular level.

Hours of discussion and brainstorming. A new start of Wellgen Medical kicks off today.

Wellgen Medical is honored to establish a formal collaborative relationship with the Faculty of Medicine, Chulalongkorn University, the most prestigious academic institution in Thailand, for the co-development of AI imaging recognition algorithms. This collaborative effort not only aims at pulmonary smear images for tuberculosis but also at cytology smear images for metastatic cancer. I am particularly happy that the academic and research experts accept Wellgen's idea of "open image" to develop AI algorithms for the good of mankind.