A Clinical Feasibility Study of a Photoacoustic Finder

-Introduction The presence of lymphatic metastases in breast cancer patients is an important prognostic factor for survival, and accurate staging leads to appropriate adjuvant treatment. Sentinel lymph node biopsy (SLNb) is a standard method used to confirm regional axillary lymphatic metastases in breast cancer patients. Sentinel lymph nodes (SLNs) are a group of initial lymph nodes (LNs) located in proximity to the tumor and connected via lymphatic vessels (LVs), hypothetically the first ones that a primary tumor drains in the regional lymphatic basin. If metastatic tumor cells are not confirmed in the excised SLNs, the incidence of morbidities, such as lymphedema, can be reduced by omitting unnecessary axillary lymph node dissection (ALND). The SLNb procedure is a dual-modal method utilizing a radioactive tracer (e.g., 99mTc) and/or blue dye to identify the SLNs. The radioactive tracer and blue dye are administered before surgery and absorbed by the lymphatic system, and eventually they flow into the SLNs. During the surgical procedures, the SLNs are identified through visual inspection of blue-dyed LVs and radioactivity detection using a gamma probe. The identified SLNs are subsequently excised and sent for pathological examination to assess the potential metastatic tumor.

The dual-modal method enhances the accuracy and efficiency of SLN identification by leveraging the distinct advantages of each method. However, the radioactive isotopes in the SLNb surgery present an inherent radiation exposure risk and necessitate specialized facilities and skilled medical personnel. These factors introduce complexities into the surgical procedure and create obstacles for the implementation of SLNb in local hospitals. Moreover, the administration of radioactive material typically requires cooperation with a nuclear medicine department, which restricts direct usage by the surgeon and may impact surgical scheduling. Lastly, because radioactive isotopes do not provide visual information, the intuitive identification of SLNs is challenging. On the other hand, blue dye visually stains the lymphatic network, enabling the intuitive identification of SLNs without radiation exposure. However, relying on visual inspection of blue-dyed SLNs may introduce inter-physician variability and potential inaccuracies in identifying SLNs, owing to variable lesion characteristics such as the presence of adipose tissue and blood. These limitations make it challenging to see blue dye within LNs, ultimately leading to reduced sensitivity in the SLN detection.

Photoacoustic (PA) imaging or sensing is a non-ionizing technique that utilizes the intrinsic light absorption properties of biological tissue components . To generate PA signals, a nanosecond pulsed laser induces repeated instantaneous thermal expansions within a sample, creating acoustic waves . These acoustic waves are then captured by an ultrasound transducer and analyzed to confirm the presence of specific constituents within the sample. PA sensing technology can detect dye-stained LNs with high sensitivity and provide a real-time quantitative representation. This method can precisely determine the presence or absence of dyed SLNs that may have been indistinguishable through visual inspection. Consequently, it can facilitate SLNb procedures without radioactive materials. In our previous studies, a cutting-edge system known as the photoacoustic finder (PAF) was successfully deviesed. It is remarkably efficacious in detecting SLNs while maintaining a high signal-to-noise ratio (SNR). The PAF combines a solid-state dye (SSD) laser handpiece and a transparent ultrasound transducer (TUT) in a precisely coaxial configuration. This study describes preclinical trials of the PAF, which confirmed its ability to accurately detect blue-dyed SLNs.

This cross-sectional clinical study was conducted ex vivo to validate the feasibility of using the PAF in a clinical setting. The process confirms the signal from excised LNs identified by the dual-modal method and the PAF before sending them to pathology. To determine its detection performance, the effectiveness of PAF is compared to the detection rate of standard SLNb. The results establish the clinical feasibility of using PAF for SLNb, providing a non-radioactive alternative.

-Study design: This study was conducted as a cross-sectional, open-label, single-arm ex vivo study within a single institution to investigate the efficacy of PAF compared to standard dual-modal methods for SLN detection in the treatment of breast cancer. The SLNb procedures followed international guidelines, using both radioisotope and blue dye mapping. SLNs were identified using gamma probe and blue dye visual inspection, then resected and labeled to reconfirm the with gamma probe and blue dye visual inspection. Subsequently, PAF was employed to capture signals from the LNs. To minimize potential errors such as labeling mistakes and LN delivery errors, the PAF system was placed in the operating room. The study enrolled women diagnosed with breast cancer who presented to the Department of Breast Surgery at St. Mary's Hospital, Seoul, The Republic of Korea.

-Sentinel lymphnode biopsy: The SLNb was performed in accordance with established international guidelines, utilizing both radioisotope and blue dye mapping methods. Five surgical oncologists participated in the study, all of whom were experienced in performing standard dual-modal SLNb and understood the clinical protocol. The radioisotope (0.1 mL of 99mTc-phytate) was administered into the subdermal lymphatic flexus under the areola within 30 minutes to 8 hours prior to operation. When the surgery started, the operator confirmed the location of the tumor. Then, a blue dye (indigo carmine, 2-5 ml) was injected peritumorally or periareolarly prior to incision, with subsequent massaging for 1 minute following injection. The axillary nodal basin was closely examined using a handheld gamma probe to detect radioactive signals and visually examined by the naked eye to detect grossly blue-dyed LNs. The identification of SLNs continued until either no further signal was detected by the gamma probe or blue-dyed LNs were no longer found in the nodal basins within the operation field. For this study, SLNs were defined as LNs with a gamma probe signal greater than 10% of the maximum signal value and/or visibly stained with blue dye. Additionally, based on the surgeon's experience, abnormally palpable LNs that were not detected by gamma probe and visual inspection were also excised, and these were labeled as non-SLNs. The SLNs and non-SLNs were further examined by PAF, and subsequently were forwarded to the pathology department for frozen section analysis.

-Sample Size: To determine the required sample size for testing the non-inferiority of PAF and the dual-modal method, the non-inferiority chi-square sample size estimator was utilized, employing a non-inferiority margin of 5%, a significance level (alpha) of 5%, and a power (1-beta) of 80%. Drawing from prior research and based on experimental data from tests, it was assumed that the visual detection rate would be approximately 78% for SLNs and 84% for PAF. Calculations revealed that a total of 157 SLNs would be necessary, corresponding to approximately 1.5 SLNs per patient, thus requiring the enrollment of 115 patients to fulfill the sample size requirements with an anticipated 10% drop-out.

-Static Analysis Methods:

The primary endpoints, which were the SLN detection rates for the gamma probe, visual inspection, and PAF were defined as follows:

Detection rate= (Number of detected SLNs)/(Total number of SLNs) ×100 [%]

The secondary endpoints encompassed the results of a non-inferiority analysis conducted using the chi-square test. Additionally, the sensitivity and specificity were assessed through ROC curve analysis as part of the additional endpoints. Regarding primary and secondary endpoints, SLNs were specifically examined, excluding non-SLNs to minimize potential surgeon subjectivity.

The Wilcoxon Mann-Whitney U-Test was performed to determine the cutoff for the PAF, and the statistical significances between 99mTc+/-, blue+/-, and non-injected groups were compared. In a non-inferiority analysis, a chi-square test was performed to compare the risk differences for 99mTc+, blue+, and PAF+ nodes. The independent samples t-test was used for other normal distributions, such as age and BMI. All statistical analyses in this study were conducted using MATLAB R2022a with the Statistics and Machine Learning Toolbox.