A diagnostic test evaluation: dynamic optical functional imaging system for early-stage breast cancer detection

Introduction

According to the findings of the Centers for Disease Control and Prevention, breast cancer is the most prevalent type of cancer and the leading cause of death among women worldwide (1,2). According to the latest global statistics, in 2020, there were 2.26 million new cases of breast cancer (24.5%) and 682,000 breast cancer-related deaths (15.5%), and in 2022, 2.3 million new cases, comprising 11.6% of all cancer cases and 666,000 breast-related deaths (6.9% of all cancer deaths) (1,2).

Due to the adverse effects of delayed diagnosis and treatment on patient prognosis (3), physicians and researchers have examined multiple methods for the early detection of breast cancer. Generally, patients with tumors smaller than 10 mm have an 85% chance of complete recovery; however, breast cancer is typically detected when the tumor is larger than 20 mm (4). According to the Institute of Medicine, the early detection of breast cancer can reduce mortality rates by enabling interventions to be implemented in the early stages of cancer development (5). Clinically, more than 90% of breast cancer patients live an additional 10 years if they are diagnosed at an early stage, and early diagnosis also enables the breasts to be preserved as much as possible. Thus, early diagnosis is a critical factor in the effective treatment of breast cancer. Consequently, there is a pressing global demand for a sophisticated, user-friendly, and non-invasive method for the early detection and diagnosis of breast cancer.

Among various breast cancer diagnostic methods, medical imaging technology, including mammography (MG), ultrasound (US), and magnetic resonance imaging (MRI), is the most widely used clinical diagnostic method (6). MG remains the universal standard for breast cancer screening and diagnosis; however, there is a growing awareness that pre-menopausal women have denser breast tissue and that MG may be less sensitive in such patients (7). US (8) has been increasingly used as an adjunctive screening and diagnostic option to MG because it is convenient, widely available, patient friendly, does not require the use of intravenous contrast media or ionizing radiation, and is relatively inexpensive. However, US has limitations in detecting breast calcifications, particularly microcalcifications. MRI and positron emission tomography (PET) are also emerging adjuncts in the diagnosis of breast cancer (9), but are still rarely employed in clinical practice. Overall, all of the above-mentioned breast imaging modalities have their structural and anatomical advantages and disadvantages, which have been summarized in the literature (9).

Currently, many research groups and companies are investigating optical breast imaging, and a number of methods and devices for breast cancer detection have been developed based on the fact that this technology can detect the functional characteristics of breast cancer (10,11). Near-infrared (NIR) spectroscopy has garnered significant research attention due to its capacity to non-invasively, conveniently, and cost effectively obtain functional information regarding the oxygenation state of the blood (12). Nioka et al. (13) measured localized metabolic signatures of cancer by NIR spectroscopy with a particular focus on the optical detection of breast cancer. Improved diffuse optical tomography (DOT) can distinguish between malignant and benign lesions by non-invasively investigating the functional, physiological, and metabolic status of breast tissue. The optical spectroscopic imaging method has been reported to detect breast cancer through angiogenesis and hypermetabolism, corresponding to chromophore concentration measurements such as oxyhemoglobin, deoxyhemoglobin (HHb), and total hemoglobin concentration (THC), oxyhemoglobin to HHb changes, and water and lipids.

To gain a comprehensive understanding of the optical and physiological paradigms, as well as the microvascular processes, a number of dynamic optical functional imaging approaches have been proposed (11,14-18), among which Dynamic Optical Breast Imaging (DOBI) (17) has recently been demonstrated to be an inexpensive, highly efficient, real-time, non-ionizing, and non-invasive method for the early diagnosis of breast cancer. The DOBI technique is based on the continuous detection and analysis of diffuse light transmission through breast tissue under pressure. This pressure modulation reveals the dynamic behavior of the optical and physiological characteristics of chromophore concentrations, representing a variety of dynamic profiles in abnormal and normal breast tissue. This preliminary clinical study sought to evaluate the effectiveness of the DOBI system in the detection of early breast cancer. We present this article in accordance with the STARD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1883/rc).

Methods

Study sample

This clinical study of the DOBI ComfortScan system was conducted at the Zhejiang Cancer Hospital in accordance with the Administrative Measures for Clinical Trials of Medical Devices (GCP). The patients enrolled in the study were aged 18 years or older and who had undergone US and for whom a breast biopsy was deemed necessary. Following the DOBI and US imaging scans, MG was performed if the patient’s Breast Imaging-Reporting and Data System (BI-RADS) classification by US imaging was ≥4A, and dynamic contrast-enhanced MRI (DCE-MRI) was performed if the patient’s BI-RADS classification by MG was ≥4A. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by Medical Ethics Committee of Zhejiang Cancer Hospital (No. IRB-2022-134) and informed consent was taken from all the patients.

The technology of the DOBI ComfortScan system

The device used for DOT in this study was the Dynamic Optical Breast Functional Imaging ComfortScan System (17,19). This non-invasive, non-ionizing medical imaging system assists physicians to detect breast cancer by providing both optical and physiological information from images of the breast. The DOBI ComfortScan system images breast tissue with neovascularization associated with tumors (i.e., tumor neovascularization) and uses breast images to determine the presence of abnormal lesions in the breast. The DOBI ComfortScan system uses diffuse light to illuminate the breast tissue from multiple angles, enabling the detection of HHb and THC, as well as changes in oxyhemoglobin to HHb over time. For more details on the DOBI ComfortScan system, please refer to (17,19).

Conventional breast imaging modalities focus on detecting static morphological (structural) features. Conversely, the DOBI ComfortScan system was designed to detect dynamic physiological changes characteristic of breast lesions (i.e., dynamic oxygen consumption due to tumor angiogenesis; that is changes in HHb). When external pressure modulation is applied to the breast, the characteristics of blood redistribution, capillary collapse, and the oxygen saturation state in the tumor region differ from those in the normal breast tissue region, and the DOBI ComfortScan system can evaluate the dynamic functional changes in THC and HHb concentration through the differences detected in light absorption (20).

Data acquisition and processing

The DOBI images shown in Figure 1A-1D were generated from three types of optical images of the breast from the DOBI ComfortScan system; that is, functional NIR (fNIR) light images, dynamic NIR (dNIR) images, and reconstructed dynamic DOT (dDOT) images. Total hemoglobin and HHb in the breast have a higher absorption of 808 and 640 nm wavelength of light, respectively. A study has shown that malignant tumors have higher THC and hemoglobin deoxygenation rates than benign tumors and normal tissues (19).

Figure 1 Images of the DOBI ComfortScan system used in this study include (A) fNIR images from three angular directions over a period; (B) dNIR images calculated from the fNIR at different times; (C) reconstructed dynamic dDOT images of the breast of three different layers at different times; and (D) different layers of DOBI images combined. dDOT, dynamic diffuse optical tomography; dNIR, dynamic near-infrared; DOBI, Dynamic Optical Breast Imaging; fNIR, functional near-infrared.

After preprocessing and normalization, the fNIR image sequence during breast compression can generate the dNIR image sequence, which can reflect the dynamic change in the HHb concentration in tumor neovascularization in the suspected lesion area as shown in Figure 1B. Changes in HHb concentration can indicate the metabolism of breast tissue. The dDOT image in Figure 1C represents the absorption coefficient matrix of optical diffusion in the breast. Finally, a DOBI image of the spatial and temporal variations of local blood hemoglobin concentration and its oxygen content (as shown in Figure 1D) is calculated from the dDOT, and the probability of breast cancer can be calculated using the positional (L), spatial (S), temporal (T), and contextual (C) features and the DOBI level reading rules.

When assessing the risk of breast cancer, the DOBI system provides a low to high probability score (0–100%) of suspected malignancy by analyzing and processing all types of images from different perspectives and using machine-learning and deep-learning algorithms.

Statistical analysis

The graphical and statistical analyses were performed using SciPy 1.7.1, Pandas 1.3.4, Scikit-learn 1.1.2, seaborn 0.11.2, and matplotlib 3.7.1, and Python 3.9.7 on anaconda3-4.0.0. For the descriptive data, the categorical or binary variables are expressed as the frequency and percentage. The accuracy (ACC), sensitivity (SEN), specificity (SPE), positive predictive value (PPV), and negative predictive value (NPV) were calculated for the DOBI system, and then compared to those of the US, MG, and DCE-MRI. For between-group comparisons, the Chi-squared test was used for the categorical variables. In all the analyses, a P value of less than 0.05 was considered statistically significant.

Results

Study sample

The data collected in this study are summarized in Table 1. The data of 320 breast lesions from 282 patients with both DOBI data and biopsy results, and the data of 183 breast lesions from 165 patients with both US and biopsy results were collected. Additionally, the data of 151 breast lesions from 136 patients with both MG and biopsy results, and the data of 74 breast lesions from 68 patients with both DCE-MRI and biopsy results were collected.

Table 2 sets out the patient data in terms of the benign/malignant status, age, and breast cup size. Of the 282 patients with both DOBI and biopsy data, 184 (65%) had benign tumors and 98 (35%) had malignant tumors. These patients had a median age of 48 years, and of the patients, 67% fell into the 40–60-year age group, 23% were younger than 40 years, and 10% were older than 60 years. The majority of the 282 patients had relatively small breast cup sizes; 87% had A- or B-cup sizes and only 13% had C- or D-cup sizes.

Table 3 shows the BI-RADS distribution for the patients enrolled in this study in terms of US, MG, and DCE-MRI. Notably, in terms of US, 42.08% of the patients had a BI-RADS classification of 1, 2, or 3, 30.6% had a BI-RADS classification of 4A, 12.02% had a BI-RADS classification of 4B, and 15.3% had a BI-RADS classification of 4C or 5. Similarly, in terms of MG, 67.55% of the patients had a BI-RADS classification of 0, 1, 2, or 3, 13.91% had a BI-RADS classification of 4A, 3.97% had a BI-RADS classification of 4B, and 14.57% had a BI-RADS classification of 4C or 5. In terms of DCE-MRI, 32.43% of the patients had a BI-RADS classification of 1, 2, or 3, 18.92% had a BI-RADS classification of 4A, 4.05% had a BI-RADS classification of 4B, and 37.83% had a BI-RADS classification of 4C or 5. The proportion of patients with a BI-RADS classification of 4C and above was higher in the DCE-MRI group than the US and MG groups because only patients with a US or MG BI-RADS classification of 4A and above underwent DCE-MRI in this study.

Breast cancer diagnosis using the DOBI system and its comparison with US, MG, and DCE-MRI

Figure 2A illustrates the performance of the DOBI ComfortScan system. In terms of the malignant and benign classification of breast lesions, the DOBI system had a high SEN cut-off value (0.32). The ACC, SEN, SPE, PPV, and NPV of the DOBI system were 81.25%, 95.96%, 74.66%, 62.91%, and 97.63%, respectively.

Figure 2 Performance of the DOBI system. (A) Performance of the DOBI system for malignant and benign breast tumor diagnosis; (B) ROC curves for the DOBI system and conventional breast imaging modalities (i.e., US, MG, and DCE-MRI). ACC, accuracy; DCE-MRI, dynamic contrast-enhanced-magnetic resonance imaging; DOBI, Dynamic Optical Breast Imaging; MG, mammography; NPV, negative predictive value; PPV, positive predictive value; ROC, receiver operating characteristic; SEN, sensitivity; SPE, specificity; US, ultrasound.

Table 4 compares the ACC, SEN, SPE, PPV, and NPV of the DOBI system, and the US and MG, using BI-RADS 4A as the cut-off value for US and MG. The results showed that the ACC, SEN, SPE, PPV, and NPV for US were 68.31%, 91.38%, 57.6%, 50%, and 93.51%, respectively, and those for MG were 83.44%, 78.57%, 85.32%, 67.35%, and 91.18%, respectively. Figure 2B shows the receiver operating characteristic (ROC) curves of the DOBI system in conjunction with those for the US, MG, and DCE-MRI breast imaging methods. As indicated by the data in Table 4, the performance of the DOBI system was found to be comparable to that of current conventional imaging methods, and thus it could be used as an adjunctive diagnostic method for breast cancer.

To further evaluate the effectiveness of the DOBI ComfortScan system in a clinical setting, we combined the results of DOBI with those of conventional imaging techniques (e.g., US and MG) for diagnostic purposes. The results of this analysis are detailed and described in the discussion section below.

Discussion

In this section, we analyze and study several cases to further explore the features and patterns of DOBI dynamic optically enabled breast images. We also describe how the DOBI system can be used to diagnose breast cancer. As Figure 3 shows, we present six cases, each comprising the following three images: an enhanced fNIR image, a DOBI image derived from a dNIR construction (i.e., a dDOT image), and a DOBI image correlation time curve. The latter depicts the change in light absorption rate in the suspicious area (SA) of the DOBI image over time. This curve represents the change in the rate of metabolism from oxygenation to HHb caused by neovascularization in the breast lesion.

Figure 3 Images of six patients. (A-F) fNIR images; (G-L) one frame of the DOBI reconstructed image sequence; (M-R) the time curves, which present the changes of HHb, at different time points of the suspicious points (indicated by red circles) in the DOBI images. DOBI, Dynamic Optical Breast Imaging; fNIR, functional near-infrared; HHb, deoxyhemoglobin.

The fNIR images reflect the absorption of 640 and 808 nm diffuse light by HHb and total hemoglobin in breast tissue over time. The concentrations of HHb and total hemoglobin varied more over time in the malignant lesions than in the benign lesions and normal tissue. The DOBI images represent the spatial and temporal characteristics of the local blood perfusion associated with the HHb concentration in the SA. The time curves represent the rate at which oxygenated hemoglobin changes to HHb in breast tissue over time, and show the difference in the temporal characteristics between the SA (blue curve) and normal breast tissue (green curve). For more details on the DOBI images of the SAs and time curves, please refer to the study by D’Aiuto et al. (17).

Before proceeding to the case-by-case analysis, let us briefly introduce the four key features of DOBI image (i.e., location, space, time, and context) that serve as the reading rules for DOBI images. The spatial images and temporal curves are shown in Figure 3G-3R. In terms of the location feature, a blue SA represents a potential malignant lesion, while a yellow SA represents a potential benign lesion. The spatial feature shows the distribution and area of the SA, representing the relative amount of change in the concentration of HHb produced in the tumor tissue with neovascularization over a given period. The temporal feature describes the metabolic profile of HHb concentration generated by tumor neovascularization in breast lesions (i.e., the blue curves in Figure 3M-3R). The contextual feature represents the difference in the metabolic profiles between the breast lesion region (i.e., the blue curves in Figure 3M-3R) and the normal region (i.e., the green curves in Figure 3M-3R). The combined analysis of the above four features leads to the DOBI imaging results, which will be expressed as the DOBI-RADS/score. Further details are provided in the case study section below.

Case studies of dynamic optical functional imaging system

In Case 1 (a 65-year-old patient with a C-cup size), the biopsy revealed an invasive ductal carcinoma with a US diagnosis of BI-RADS 4B, and a DCE-MRI diagnosis of BI-RADS 4C (Figure 3A). In Case 2 (a 70-year-old patient with an A-cup size), the biopsy revealed an invasive breast cancer, with a US diagnosis of BI-RADS 4C (Figure 3A). These patients had DOBI system diagnosis scores of 93 and 88, respectively, indicating malignant diagnoses. The DOBI images in Figure 3G,3H showed location features characterized by the presence of blue SAs, which suggested the possible presence of malignant tumors. The area was located in the center of the breast in Case 1, and in the left center of the breast in Case 2. In both cases, the blue areas were darkly shaded, uneven in intensity, and had irregular or rough borders. This might be due to the large number of neovascularized capillaries surrounding the tumors. These capillaries increase and accumulate deoxyhemoglobin during compression, which further increases light absorption in the tissue. In terms of the spatial characteristics of the DOBI images, Case 1 showed focal, peaked features, while Case 2 showed non-peaked features, and both showed stable blue regions. A comparison of the SA curves of the DOBI images and the curves of the normal tissues in Figure 3M,3N revealed that the temporal curves of the suspected abnormal regions showed a decreasing trend, which differed from the curves of the metabolically normal regions.

Tumor neovascularization has a higher metabolic rate from oxyhemoglobin to deoxyhemoglobin, and exhibits spatial and temporal features associated with malignancy in DOBI images. According to the DOBI reading rules, and based on the four features of the DOBI images, the final calculated DOBI scores showed malignancy degrees of 93 for Case 1 and 88 for Case 2, respectively. After evaluating the fNIR and DOBI images of Cases 1 and 2, the DOBI system output the diagnoses of these two cases as malignant.

As Figure 3C,3D showed, Case 3 (a 53-year-old with an A-cup size) and Case 4 (a 37-year-old with an A-cup size) had benign lesions as indicated by their biopsy results. The breast US diagnosis for Case 3 was BI-RADS 4A, while the MG diagnosis for Case 4 was BI-RADS 3. The DOBI scores for Cases 3 and 4 were 44 and 52, respectively, indicating benign diagnoses. As Figure 3C,3D showed, the suspected abnormal area in Case 3 was located in the top center of the breast, while that in Case 4 was located in the center of the breast. Both suspected areas exhibited slightly low intensity and relatively smooth borders.

As Figure 3I,3J showed, in terms of the location characteristics, the DOBI images for Cases 3 and 4 showed no blue SAs, but showed yellow areas, indicating possible benign characteristics. No blue SA was found in either Case 3 or 4; thus, no spatial features indicative of malignancy were observed in either case.

We then examined the SA curves of the DOBI images in Figure 3O,3P. The blue curves of the suspected abnormal region in Cases 3 and 4 showed initial waviness, ultimately displaying a pronounced upward trend, which suggested that the blue SA regions were likely benign. The green curves of the normal tissue and the blue curves of the DOBI images in Figure 3O exhibited similar characteristics, indicating that the metabolic characteristics of deoxyhemoglobin in the normal tissue and the SAs in the breast tissue were comparable.

The final DOBI score was determined by assessing the four features of the DOBI images in accordance with the established reading rules. The results indicated that both cases were benign; Case 3 had a DOBI score of 44, while Case 4 had a DOBI score of 52. After evaluating the fNIR and the DOBI images of the two cases, the DOBI system output a diagnosis with a probability of malignancy of 0.21 for Case 3 and a diagnosis with a probability of malignancy of 0.15 for Case 4. The DOBI system output the diagnoses of these two cases as benign.

Figure 3E,3F presented the results for Case 5 (a 41-year-old with an A-cup size) and Case 6 (a 53-year-old with a B-cup size). The biopsy of Case 5 revealed a fibroadenoma, with US findings of BI-RADS 4A, and DCE-MRI findings of BI-RADS 2. The biopsy of Case 6 revealed an invasive carcinoma, for which no morphologic imaging results were available. As Figure 3E,3F showed, the suspected abnormal areas in Cases 5 and 6 were located in the center of the breast, exhibiting dark shadows and irregular or rough borders.

Figure 3K,3L showed the DOBI spatial characteristics. The spatial features of the blue regions in both Cases 5 and 6 exhibited diffuse, flat-bottomed, and wandering patterns, which suggested that the SAs were not well correlated with the malignant tumor type. However, no overt benign tumor characteristics were observed.

Finally, we examined the SAs and normal curves of the DOBI images in Figure 3Q,3R. The time curves of the suspected abnormal regions displayed a downward trend, with fluctuations in the blue line, and slightly different characteristics compared to the curves of the normal regions. Consequently, the temporal and contrast features of these DOBI images showed indeterminate positive or negative characteristic behaviors. Based on the established reading rules and the integration of these four features, the DOBI score results indicated indeterminate outcomes for these two cases; Case 5 had a score of 73, while Case 6 had a score of 64. After evaluating the fNIR and the DOBI images, the DOBI system output the diagnoses of these indeterminate cases as having probabilities of malignancy of 0.45 and 0.57, respectively.

The clinical effectiveness of morphological and functional combinations using the DOBI system with conventional imaging methods

Based on the preliminary clinical results, the DOBI system had a high SEN and a reasonable SPE. As a new dynamic functional imaging method, the DOBI system has the potential to be used as a first-line detection method for breast cancer, including for breast cancer risk assessment. The current study also showed that combining the DOBI dynamic functional system with conventional imaging techniques (e.g., US or MG) can improve the clinical diagnostic ACC of early breast cancer.

To minimize the number of false-negative (FN) cases, one clinical application integrated the DOBI model (cut-off =0.32) with US results classified as BI-RADS 3 and below. As Figure 4 shows, there were a total of 76 cases that had both biopsy and DOBI results, as well as US results of BI-RADS 1, 2, and 3. All 76 of these cases were considered benign based on the US test. The biopsy results indicated that five of these 76 cases were malignant; thus, there would have been five FN results if only US had been used. For these cases, we employed a combination of DOBI and US. The DOBI system was used to assess a total of 76 cases, of which 19 were classified as malignant by the DOBI system. This included all five cases that were initially determined to be malignant using US alone. Consequently, the number of FN cases decreased from five to zero after integrating the DOBI System with the US test. However, it also provided a malignant diagnosis in 14 cases that were actually benign.

Figure 4 Clinical application of using the DOBI dynamic functional system to reduce the FN rate for breast cancer detection. BI-RADS, Breast Imaging-Reporting and Data System; DOBI, Dynamic Optical Breast Imaging; FN, false-negative; MG, mammography; US, ultrasound.

As Figure 4 shows, a total of 101 cases had both biopsy and DOBI results, as well as MG results of BI-RADS 1, 2, or 3. All 101 cases were considered benign based on the MG test. The biopsy results indicated that 9 of these 101 cases were determined to be malignant; thus, there would have been 9 FNs if only MG had been used. When the DOBI model (cut-off value =0.32) was combined with MG, 32 of the 101 cases were identified as malignant by the DOBI system. This included all nine cases that were determined to be malignant. Consequently, the number of FN cases decreased from nine to zero after integrating the DOBI system with the MG test. However, it also provided a malignant diagnosis in 23 cases that were actually benign.

The DOBI system can also be applied to reduce the false-positive (FP) rate in breast cancer detection, where conventional imaging (US and/or MG) results are BI-RADS 4A or 4B. As Figure 5 shows, there were 77 cases with US results of BI-RADS 4A or 4B, and those 77 cases were considered malignant based on US testing. Of the 77 cases with US results of BI-RADS 4A or 4B, the biopsy results indicated that 52 of the 77 cases were benign; thus, there were 52 FP cases. After combining the diagnosis with the DOBI model (cut-off value =0.32), 38 of the above 77 positive cases were determined to be benign. Among the 38 cases, there was only one case whose biopsy was malignant. Consequently, 37 cases were re-categorized from FP to true-negative (TN). As a result, the number of FP cases was reduced from 52 to 15.

Figure 5 Clinical application of using DOBI dynamic functional system to reduce the FP rate of breast cancer detection at early-stage while US and MG diagnosed them as malignant. BI-RADS, Breast Imaging-Reporting and Data System; DOBI, Dynamic Optical Breast Imaging; FN, false-negative; FP, false-positive; MG, mammography; US, ultrasound.

According to the data in Figure 5, 27 cases were classified as BI-RADS 4A or 4B, and MG assessed these cases as malignant. However, the biopsy results revealed that 16 of the 27 cases were benign; thus, there were 16 FP cases. When a DOBI model cut-off value of 0.32 was applied, 14 of the 27 FP cases were identified as benign, and the biopsy results of these 14 cases showed that there were 13 TN cases and one malignant case. Consequently, by integrating the DOBI system with the MG results, the number of FP cases identified by the MG (BI-RADS 4A or 4B) test was significantly reduced from 16 to 3. Based on the analysis of the aforementioned clinical data, the DOBI system can be used in conjunction with US or MG, depending on the requirements, to minimize the number of FN or FP cases in the diagnosis of breast cancer.

While DOBI’s lower cost and non-ionizing nature are advantages over DCE-MRI, future studies should include formal cost-effectiveness analyses and larger cohorts to validate its economic and clinical utility.

Limitations

This study did not evaluate DOBI’s performance in dense breast tissue subgroups, though its functional imaging mechanism may offer advantages over MG in this population. This study did not evaluate DOBI’s performance across specific breast cancer subtypes (e.g., triple-negative or HER2-positive), which may exhibit distinct angiogenic or metabolic profiles. This preliminary analysis did not investigate long-term outcomes (e.g., survival benefits or stage-specific intervention rates) associated with DOBI-assisted diagnosis.

Conclusions

The DOBI ComfortScan system employs a dynamic functional imaging approach to assess the volume of blood hemoglobin concentration and its rate of change from oxyhemoglobin to HHb in the initial neovascularization of breast tumors from multiple angles using diffuse light. The system then analyzes the differences in the specific spectra’s absorption under pressure modulation to distinguish between normal tissue, benign lesions, and malignant lesions in breast tissue. The study found that the DOBI system had a strong diagnostic ACC of 81.25%, a SEN of 95.96%, a SPE of 74.66%, a PPV of 62.91%, and a NPV of 97.63%. These results indicate that the DOBI system is comparable to MG and better than US in this dataset in diagnosing breast cancer.

The DOBI ComfortScan system had a high SEN and a high NPV in this study, suggesting its potential as a valuable tool for early breast cancer screening. In clinical settings, the DOBI ComfortScan system could be used in conjunction with US or MG, as needed, to reduce the FN rates for US or MG diagnoses of BI-RADS 3 and below, and the FP rates for US or MG diagnoses of BI-RADS 4A and/or 4B.

In summary, the DOBI ComfortScan system is a painless, non-invasive, non-radiologic, dynamic optical breast functional imaging device that is safe, convenient, comfortable, and cost effective. This study included a limited number of cases, but the preliminary results are both encouraging and promising. Further research on the DOBI ComfortScan system will be conducted to gather more substantial data.

Acknowledgments

The authors would like to thank all parties and participants of the clinical trial on the DOBI ComfortScan system, conducted at the Zhejiang Cancer Hospital. We would like to especially thank Prof. Xiguo Yuan, Prof. Qiang Yu, Dr. Xue Li, and Dr. Baiyan Zhang at Xidian University Hangzhou Institute of Technology for the development of the breast cancer classification algorithm. Last but not least, we would like to thank Luyan Liu, Yan Gu, Ying Li, and Lei Jiang at DOBI Medical who conducted the statistical analysis and prepared the report.

Footnote

Reporting Checklist: The authors have completed the STARD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1883/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1883/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1883/prf

Funding: This study was supported by the Zhejiang Province Science and Technology Project of Medicine and Health (Nos. 2022KY632 and 2021KY085).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1883/coif). J.Z. is from DOBI Medical, Shirley, MA, USA. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by Medical Ethics Committee of Zhejiang Cancer Hospital (No. IRB-2022-134) and informed consent was taken from all the patients.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-49. [Crossref] [PubMed]
2. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
3. Huo Q, Cai C, Zhang Y, et al. Delay in diagnosis and treatment of symptomatic breast cancer in China. Ann Surg Oncol 2015;22:883-8. [Crossref] [PubMed]
4. Roslidar R, Rahman A, Muharar R, et al. A review on recent progress in thermal imaging and deep learning approaches for breast cancer detection. IEEE 2020;8:116176-94.
5. Institute of Medicine (US) and National Research Council (US) Committee on the Early Detection of Breast Cancer; Patlak M, Nass SJ, et al., editors. Mammography and Beyond: Developing Technologies for the Early Detection of Breast Cancer: A Non-Technical Summary. Washington (DC): National Academies Press (US); 2001.
6. Jafari SH, Saadatpour Z, Salmaninejad A, et al. Breast cancer diagnosis: Imaging techniques and biochemical markers. J Cell Physiol 2018;233:5200-13. [Crossref] [PubMed]
7. Breast Cancer Screening GuidelinesUSPSTF. Available online: https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/breast-cancer-screening
8. Geisel J, Raghu M, Hooley R. The Role of Ultrasound in Breast Cancer Screening: The Case for and Against Ultrasound. Semin Ultrasound CT MR 2018;39:25-34. [Crossref] [PubMed]
9. Alshehri A, AlSaeed D. Breast cancer diagnosis using artificial intelligence approaches: A systematic literature review. Intelligent Automation & Soft Computing 2023;37:939-70.
10. Liu H, Gu Y, Kim JG, et al. Near-infrared spectroscopy and imaging of tumor vascular oxygenation. Methods Enzymol 2004;386:349-78. [Crossref] [PubMed]
11. Grosenick D, Rinneberg H, Cubeddu R, et al. Review of optical breast imaging and spectroscopy. J Biomed Opt 2016;21:091311. [Crossref] [PubMed]
12. Paulsen KD, Meaney PM, Gilman L, et al. Near infrared spectroscopic imaging: Translation to clinic. Alternative Breast Imaging: Four Model-Based Approaches; 2005:201-26.
13. Nioka S, Chance B. NIR spectroscopic detection of breast cancer. Technol Cancer Res Treat 2005;4:497-512. [Crossref] [PubMed]
14. Barbour RL, Schmitz CH, Graber HL, et al. Dynamic functional imaging of the healthy and cancerous breast by optical tomography. In: Biomedical Topical Meeting, Optica Publishing Group; 2002:TuB2.
15. Barbour RL, Graber HL, Pei Y, et al. Optical tomographic imaging of dynamic features of dense-scattering media. J Opt Soc Am A Opt Image Sci Vis 2001;18:3018-36. [Crossref] [PubMed]
16. Athanasiou A, Vanel D, Balleyguier C, et al. Dynamic optical breast imaging: a new technique to visualise breast vessels: comparison with breast MRI and preliminary results. Eur J Radiol 2005;54:72-9. [Crossref] [PubMed]
17. D'Aiuto M, Frasci G, Barretta ML, et al. The dynamic optical breast imaging in the preoperative workflow of women with suspicious or malignant breast lesions: development of a new comprehensive score. ISRN Oncol 2012;2012:631917. [Crossref] [PubMed]
18. Shanthakumar D, Leiloglou M, Kelliher C, et al. A Comparison of Spectroscopy and Imaging Techniques Utilizing Spectrally Resolved Diffusely Reflected Light for Intraoperative Margin Assessment in Breast-Conserving Surgery: A Systematic Review and Meta-Analysis. Cancers (Basel) 2023;15:2884. [Crossref] [PubMed]
19. Petrillo A, Catalano O, Fusco R, et al. Optical imaging of the breast: evaluation of deoxyhemoglobin concentration alteration in 166 patients with suspicious breast lesions. Eur Radiol Exp 2018;2:8. [Crossref] [PubMed]
20. McDonald DM, Choyke PL. Imaging of angiogenesis: from microscope to clinic. Nat Med 2003;9:713-25. [Crossref] [PubMed]

(English Language Editor: L. Huleatt)