Proof of concept for the use of trained sniffer dogs to detect osteosarcoma.

Sarcomas are mesenchymal cancers which often show an aggressive behavior and patient survival largely depends on an early detection. In last years, much attention has been given to the fact that cancer patients release specific odorous volatile organic compounds (VOCs) that can be efficiently detected by properly trained sniffer dogs. Here, we have evaluated for the first time the ability of sniffer dogs (n = 2) to detect osteosarcoma cell cultures and patient samples. One of the two dogs was successfully trained to discriminate osteosarcoma patient-derived primary cells from mesenchymal stem/stromal cells (MSCs) obtained from healthy individuals. After the training phase, the dog was able to detect osteosarcoma specific odor cues in a different panel of 6 osteosarcoma cell lines with sensitivity and specificity rates between 95 and 100%. Moreover, the same VOCs were also detected by the sniffer dog in saliva samples from osteosarcoma patients (n = 2) and discriminated from samples from healthy individuals with a similar efficacy. Altogether, these results indicate that there are common odor profiles shared by cultures of osteosarcoma cells and body fluid samples from patients and provide a first proof of concept about the potential of canine odor detection as a non-invasive screening method to detect osteosarcomas.

Robustifying the Deployment of tinyML Models for Autonomous Mini-Vehicles.

Standard-sized autonomous vehicles have rapidly improved thanks to the breakthroughs of deep learning. However, scaling autonomous driving to mini-vehicles poses several challenges due to their limited on-board storage and computing capabilities. Moreover, autonomous systems lack robustness when deployed in dynamic environments where the underlying distribution is different from the distribution learned during training. To address these challenges, we propose a closed-loop learning flow for autonomous driving mini-vehicles that includes the target deployment environment in-the-loop. We leverage a family of compact and high-throughput tinyCNNs to control the mini-vehicle that learn by imitating a computer vision algorithm, i.e., the expert, in the target environment. Thus, the tinyCNNs, having only access to an on-board fast-rate linear camera, gain robustness to lighting conditions and improve over time. Moreover, we introduce an online predictor that can choose between different tinyCNN models at runtime-trading accuracy and latency-which minimises the inference's energy consumption by up to 3.2×. Finally, we leverage GAP8, a parallel ultra-low-power RISC-V-based micro-controller unit (MCU), to meet the real-time inference requirements. When running the family of tinyCNNs, our solution running on GAP8 outperforms any other implementation on the STM32L4 and NXP k64f (traditional single-core MCUs), reducing the latency by over 13× and the energy consumption by 92%.