This study was successful in imaging high quantitative values of fecal occult blood on the stool surface through HSI. To the best of our knowledge, this technology is the first of its kind in the world, and we are confident that this will have a significant impact on the development of medical devices in the future.

In the United States, the mortality rate of colorectal cancer has been declining since around 1980, largely attributed to screening tests such as fecal occult blood tests, sigmoidoscopy, and total colonoscopy, which are widely undergone by many United States citizens [18,19,20]. The effectiveness of colorectal cancer screening in preventing the disease is well recognized, and FIT-based colorectal cancer screening programs have been implemented in various countries [21]. FIT, a low-cost, noninvasive test, has been shown to correlate with lower colorectal cancer mortality rates despite having more false-negative results than colonoscopy [22, 23]. A meta-analysis in the United States and Europe demonstrated that fecal occult blood tests over a 2-day period have a sensitivity of 77% and specificity of 93% for detecting colorectal cancer [3]. In recent years, the development of new detection modalities for malignant diseases has been reported. These include liquid biopsy, which utilizes body fluid samples such as blood, urine, and stool [24, 25], and a novel method using N-NOSE [26]. However, none of these methods has provided data surpassing the utility of fecal occult blood testing in detecting colorectal cancer. Recently, it has been reported that a next-generation stool test kit, which detects trace amounts of colorectal cancer DNA, can identify 93% of colorectal cancers in a single fecal test [27]. Nevertheless, this test is similar to other tests that require the intention to undergo a checkup and is unlikely to reduce the number of people who remain unexamined.

However, in Japan, where the 2-day FIT method has been used for colorectal cancer screening since 1997, the mortality rate remains higher than those in Europe and the United States. In addition, the low screening rate has become a concern. Possible reasons for this may include (1) the lack of compulsion to take the test, (2) a population with low interest in health care, and (3) insufficient widespread understanding of the test's usefulness. Furthermore, the fecal occult blood test is time-consuming as it requires a sample to be scraped from the stool and necessitates innovative measures to increase the examination rate. Therefore, as the development of an imaging technology capable of rapidly and easily identifying occult blood in stools would address this issue, we explored detecting blood in stool using HSI. We hypothesized that if implemented, fecal occult blood could be instantly measured during routine bowel movements by integrating this technology into toilets.

Setting cutoff values is important in constructing an algorithm to discriminate blood in stool using HSI. The quantitative measurement of fecal occult blood for colorectal cancer screening has been extensively reported, and fecal occult blood quantitative values are particularly high when the cancer is at the advanced stage [10, 11]. In this study, we aimed to develop a toilet that would increase the number of people who receive screening and detect as many colorectal cancers as possible among those who do not receive screening. Therefore, instead of the commonly used screening cutoff value of 100 ng/ml [28,29,30], we set a higher cutoff value that would detect lesions at a high rate in daily bowel movements and have a low false positive rate. The sensitivity of colorectal cancer and adenoma lesions was compared, but no significant difference was found. Therefore, we utilized the 400 ng/ml cutoff value, assumed to yield fewer false negatives, for constructing the algorithm. The algorithm's precision was confirmed with impressive results: 90.0% discrimination accuracy for each analyzed region and 92.0% discrimination accuracy per case, indicating that HSI can discriminate high quantitative values of occult blood in stool, with high accuracy. Validation of variation further demonstrated high discrimination accuracy of 90.8% (95% confidence interval: 86.5–94.1%) by region and 88.0% (95% confidence interval: 75.7–95.5%) by case. High sensitivity and specificity were also observed. Furthermore, many of the false positive cases were also cancer cases, suggesting that if this discrimination image is positive, colon cancer would be detected at a high rate. In the future, we would like to continue our research and further improve analytical techniques so that I can image the blood at a cutoff value of 100 ng/ml, which is the level used in screening tests.

The advantage of HSI is that it can measure the entire stool. The images could be classified into three major categories: whole positive pattern, partial positive pattern, and whole negative pattern. Many of the results were partial positive patterns, suggesting that there may be an uneven distribution of fecal occult blood values even on the stool surface, which is conceivably a factor increasing the false-negative rate of conventional fecal occult blood tests where the participant collects the specimen.

In this study, blood was mixed with various liquids (water, milk, tomato juice, tea, and coffee) to verify whether HSI can specifically recognize Hb and not merely color, photographing only blood in all cases. Particularly with milk, even though the blood had settled and some parts of the blood serum were not visible to the naked eye, they could be depicted in the discriminant image. This indicates that Hb is identified in HSI not by color difference but as a characteristic of reflected and absorbed light wavelengths. Moreover, the ability to detect it even when passing through liquid suggests that detection in the toilet might also be feasible. However, in experiments using a stool model, while blood on the stool surface was detectable, blood within the stool remained unrecognized, suggesting to a potential limitation of the imaging technology. In the future, we plan to compare the detection ability of lesions in an RCT after creating a practical device to determine whether the inability to detect fecal occult blood in stool will be a limitation to the practical application of the device.

Though there have been numerous toilet studies conducted in recent years, such as those monitoring health by analyzing stool properties according to the Bristol Stool Scale [31] and those attempting to implement a system to indirectly measure blood glucose levels in urine using a color sensor [32], no studies have reported implementing technology to analyze blood in stool for colorectal cancer screening. This study suggests that detecting fecal occult blood in the stool during daily bowel movements could promptly inform patients about the presence of fecal occult blood on the stool surface without delay. Moreover, the ability to analyze the entire stool surface is expected to enhance accuracy compared to traditional fecal occult blood tests. Furthermore, the development of a toilet that alerts relevant parties in real-time to the need for further examination and encourages them to undergo screening or visit a medical institution has begun, aiming to facilitate early detection of colorectal cancer in individuals who have little interest in health care and do not participate in screening. If this toilet is developed, we plan to install it first in public facilities such as hotels and shopping centers. If many people with colorectal cancer are identified through the use of this toilet, it is expected that use of this toilet will increase, leading to an increase in the screening rate and ultimately a decrease in the colorectal cancer mortality rate.

The limitations of this study are as follows. First, since many of the study participants had cancer, only a few of them had no findings. Therefore, a large-scale prospective study is required for further validation of findings. Second, the imaging conditions used in the analysis involved photographs obtained under dark room conditions, not underwater. Additional studies are needed for clinical application to toilets, considering halation and light reflection during image processing when the feces are submerged in water.

This study demonstrates that HSI can facilitate the imaging of high quantitative values of fecal occult blood on stool surfaces, highlighting its potential for future clinical applications in toilets to aid in the prevention of colon cancer.