Background: Current artificial intelligence (AI) in histopathology typically specializes on a single task, resulting in a heavy workload of collecting and labeling a sufficient number of images for each type of cancer. Heterogeneous transfer learning (HTL) is expected to alleviate the data bottlenecks and establish models with performance comparable to supervised learning (SL). Methods: An accurate source domain model was trained using 28,634 colorectal patches. Additionally, 1000 sentinel lymph node patches and 1008 breast patches were used to train two target domain models. The feature distribution difference between sentinel lymph node metastasis or breast cancer and CRC was reduced by heterogeneous domain adaptation, and the maximum mean difference between subdomains was used for knowledge transfer to achieve accurate classification across multiple cancers. Result: HTL on 1000 sentinel lymph node patches (L-HTL-1000) outperforms SL on 1000 sentinel lymph node patches (L-SL-1-1000) (average area under the curve (AUC) and standard deviation of L-HTL-1000 vs. L-SL-1-1000: 0.949 ± 0.004 vs. 0.931 ± 0.008, p value = 0.008). There is no significant difference between L-HTL-1000 and SL on 7104 patches (L-SL-2-7104) (0.949 ± 0.004 vs. 0.948 ± 0.008, p value = 0.742). Similar results are observed for breast cancer. B-HTL-1008 vs. B-SL-1-1008: 0.962 ± 0.017 vs. 0.943 ± 0.018, p value = 0.008; B-HTL-1008 vs. B-SL-2-5232: 0.962 ± 0.017 vs. 0.951 ± 0.023, p value = 0.148. Conclusions: HTL is capable of building accurate AI models for similar cancers using a small amount of data based on a large dataset for a certain type of cancer. HTL holds great promise for accelerating the development of AI in histopathology.

Melanoma is a common, highly fatal skin cancer. Photoacoustic imaging can achieve highly sensitive and high-contrast detection of melanin molecules in tissues, also inheriting the high penetration depth and high spatial resolution characteristics of ultrasound imaging, thus it is a very promising non-invasive diagnostic tool for early melanoma. In this work, we built an acoustic-resolution-based photoacoustic microscopy system, using 1064 nm/532 nm pulsed light to observe melanoma in the back of a mouse with simultaneous photoacoustic/ultrasound imaging. Through the fusion of multi-modal images, accurate positioning of melanoma and its surrounding normal tissues were realized. This work will further promote the application of photoacoustic imaging in the clinical diagnosis of early melanoma.

Neurovascular coupling in epilepsy is poorly understood; its study requires simultaneous monitoring of hemodynamic changes and neural activity in the brain. Here for the first time we present a combined real-time 3D photoacoustic tomography (PAT) and electrophysiology/electroencephalography (EEG) system for the study of neurovascular coupling in epilepsy, whose ability was demonstrated with a pentylenetetrazol (PTZ) induced generalized seizure model in rats. Two groups of experiments were carried out with different wavelengths to detect the changes of oxy-hemoglobin (HbO2) and deoxy-hemoglobin (HbR) signals in the rat brain. We extracted the average PAT signals of the superior sagittal sinus (SSS), and compared them with the EEG signal. Results showed that the seizure process can be divided into three stages. A 'dip' lasting for 1-2 min in the first stage and the following hyperfusion in the second stage were observed. The HbO2 signal and the HbR signal were generally negatively correlated. The change of blood flow was also estimated. All the acquired results here were in accordance with other published results. Compared to other existing functional neuroimaging tools, the method proposed here enables reliable tracking of hemodynamic signal with both high spatial and high temporal resolution in 3D, so it is more suitable for neurovascular coupling study of epilepsy.

While photoacoustic computed tomography (PACT) is generally built with planar transducers of finite size, most current reconstruction algorithms assume the transducer to be an ideal point, which leads to a spinning blur in the consequently obtained PACT images due to the model mismatch. In this work, we put forward an improved back-projection method that factors in the geometry of the transducers to improve the tangential resolution for the reconstruction of 2D circular-scanning-based photoacoustic tomography. Extensive simulations and experiments were carried out to study the adaptability and stability of the proposed method. Results show that this method can effectively restore the tangential distortion of the PACT image for both simulated and experimental data. Results indicated that the improvement of the tangential resolution is more obvious for transducers with larger size. We also demonstrated the application of this method to transducers other than planar, and results show that the reconstructed image quality can be significantly affected by the shape and position of the transducers used. This study may help to guide the selection of transducer and design of scanning strategy in PACT.

The inadequacy of existing colorectal imaging tools has significantly obstructed the efficient detection of colorectal cancer. To address this issue, this work presents the cross-scale endoscopic imaging of rectal tumors with a combined photoacoustic/ultrasound tomography system and wide-field optical microscopy. This multimodal system combines the merits of centimeter-scale deep penetration, multi-spectral imaging, cross-scale imaging ability, low system cost, and 360° view in a single modality. Results indicated that the proposed system could reliably depict the location of the cancer invasion depth spectroscopically with indocyanine green The tumor angiogenesis can be well identified in the wide-field optical imaging mode, which helps to localize the tumors and guide the following photoacoustic/ultrasound scan. This work may facilitate the accurate characterization of colorectal cancer and promote the clinical translation of photoacoustic-based colorectal endoscopy.

Existing acoustic-resolution photoacoustic/ultrasonic endoscopy (PA/USE) generally employs a point-focused transducer for ultrasound detection, which is only sensitive in its focal region, thus the lateral resolution and sensitivity drop dramatically when the targets move far from its focus. Even if a dynamic focusing algorithm is applied, the sensitivity out of the transducer focus is still much lower than that in the focus in ultrasonic imaging mode. In this work, we propose an acoustic-resolution PA/USE with a line-focused transducer to realize automatic focusing for the first time. In comparison to a point-focused transducer, the line-focused transducer emits a more uniform sound field, causing the original signal intensity and signal-to-noise ratio (SNR) of the front and rear targets to be closer in the radial direction, which is beneficial for improving target signal uniformity in ultrasonic imaging. Simultaneously, we improved the resolution of the defocus area by modifying a prior work of back-projection (BP) reconstruction algorithm typically used in point-focused transducer based PAE and applying it to line-focused PA/USE. This combined approach may significantly enhance the depth of field of ultrasonic imaging and the resolution of the defocus zone in PA/US imaging, compared to the conventional method. Sufficient numerical simulations and phantom experiments were performed to verify this method. The results show that our method can effectively improve the lateral resolution in the image’s defocused region to achieve automatic focusing and perfectly solve the defect of the target signal difference in the far-focus region in ultrasonic imaging, while also enhancing the image SNR and contrast. The proposed method in this paper lays foundations for the realization of photoacoustic/ultrasonic combined endoscopy with enhanced lateral resolution and depth of field, which can potentially benefit a many of biomedical applications.