Fast-track development and multi-institutional clinical validation of an artificial intelligence algorithm for detection of lymph node metastasis in colorectal cancer

Lymph node metastasis (LNM) detection can be automated using artificial intelligence-based diagnostic tools. Only limited studies have addressed this task for colorectal cancer. The aim of this study was to develop of a clinical-grade digital pathology tool for LNM detection in colorectal cancer (CRC) using the original fast-track framework. The training cohort included 432 slides from one department. A segmentation algorithm detecting 8 relevant tissue classes was trained. The test cohorts consisted of materials from five pathology departments digitized by four different scanning systems. A high-quality, large training dataset was generated within 7 days, and a minimal amount of annotation work using fast-track principles. The AI tool showed very high accuracy for LNM detection in all cohorts, with sensitivity, negative predictive value, and specificity ranges of 0.980-1.000, 0.997-1.000, and 0.913-0.990, correspondingly. Only 5 of 14460 analyzed test slides with tumor cells over all cohorts were classified as false negative (3/5 representing clusters of tumor cells in lymphatic vessels). A clinical-grade tool was trained in a short time using fast-track development principles and validated using the largest international, multi-institutional, multi-scanner cohort of cases to date, showing very high precision for LNM detection in CRC. We are releasing a part of the test datasets to facilitate academic research.

Development and Evaluation of Machine Learning in Whole-Body Magnetic Resonance Imaging for Detecting Metastases in Patients With Lung or Colon Cancer: A Diagnostic Test Accuracy Study

OBJECTIVES

Whole-body magnetic resonance imaging (WB-MRI) has been demonstrated to be efficient and cost-effective for cancer staging. The study aim was to develop a machine learning (ML) algorithm to improve radiologists' sensitivity and specificity for metastasis detection and reduce reading times.

MATERIALS AND METHODS

A retrospective analysis of 438 prospectively collected WB-MRI scans from multicenter Streamline studies (February 2013-September 2016) was undertaken. Disease sites were manually labeled using Streamline reference standard. Whole-body MRI scans were randomly allocated to training and testing sets. A model for malignant lesion detection was developed based on convolutional neural networks and a 2-stage training strategy. The final algorithm generated lesion probability heat maps. Using a concurrent reader paradigm, 25 radiologists (18 experienced, 7 inexperienced in WB-/MRI) were randomly allocated WB-MRI scans with or without ML support to detect malignant lesions over 2 or 3 reading rounds. Reads were undertaken in the setting of a diagnostic radiology reading room between November 2019 and March 2020. Reading times were recorded by a scribe. Prespecified analysis included sensitivity, specificity, interobserver agreement, and reading time of radiology readers to detect metastases with or without ML support. Reader performance for detection of the primary tumor was also evaluated.

RESULTS

Four hundred thirty-three evaluable WB-MRI scans were allocated to algorithm training (245) or radiology testing (50 patients with metastases, from primary 117 colon [n = 117] or lung [n = 71] cancer). Among a total 562 reads by experienced radiologists over 2 reading rounds, per-patient specificity was 86.2% (ML) and 87.7% (non-ML) (-1.5% difference; 95% confidence interval [CI], -6.4%, 3.5%; P = 0.39). Sensitivity was 66.0% (ML) and 70.0% (non-ML) (-4.0% difference; 95% CI, -13.5%, 5.5%; P = 0.344). Among 161 reads by inexperienced readers, per-patient specificity in both groups was 76.3% (0% difference; 95% CI, -15.0%, 15.0%; P = 0.613), with sensitivity of 73.3% (ML) and 60.0% (non-ML) (13.3% difference; 95% CI, -7.9%, 34.5%; P = 0.313). Per-site specificity was high (>90%) for all metastatic sites and experience levels. There was high sensitivity for the detection of primary tumors (lung cancer detection rate of 98.6% with and without ML [0.0% difference; 95% CI, -2.0%, 2.0%; P = 1.00], colon cancer detection rate of 89.0% with and 90.6% without ML [-1.7% difference; 95% CI, -5.6%, 2.2%; P = 0.65]). When combining all reads from rounds 1 and 2, reading times fell by 6.2% (95% CI, -22.8%, 10.0%) when using ML. Round 2 read-times fell by 32% (95% CI, 20.8%, 42.8%) compared with round 1. Within round 2, there was a significant decrease in read-time when using ML support, estimated as 286 seconds (or 11%) quicker ( P = 0.0281), using regression analysis to account for reader experience, read round, and tumor type. Interobserver variance suggests moderate agreement, Cohen κ = 0.64; 95% CI, 0.47, 0.81 (with ML), and Cohen κ = 0.66; 95% CI, 0.47, 0.81 (without ML).

CONCLUSIONS

There was no evidence of a significant difference in per-patient sensitivity and specificity for detecting metastases or the primary tumor using concurrent ML compared with standard WB-MRI. Radiology read-times with or without ML support fell for round 2 reads compared with round 1, suggesting that readers familiarized themselves with the study reading method. During the second reading round, there was a significant reduction in reading time when using ML support.

Computer-Assisted Diagnosis of Lymph Node Metastases in Colorectal Cancers Using Transfer Learning With an Ensemble Model

Screening of lymph node metastases in colorectal cancer (CRC) can be a cumbersome task, but it is amenable to artificial intelligence (AI)-assisted diagnostic solution. Here, we propose a deep learning-based workflow for the evaluation of CRC lymph node metastases from digitized hematoxylin and eosin-stained sections. A segmentation model was trained on 100 whole-slide images (WSIs). It achieved a Matthews correlation coefficient of 0.86 (±0.154) and an acceptable Hausdorff distance of 135.59 μm (±72.14 μm), indicating a high congruence with the ground truth. For metastasis detection, 2 models (Xception and Vision Transformer) were independently trained first on a patch-based breast cancer lymph node data set and were then fine-tuned using the CRC data set. After fine-tuning, the ensemble model showed significant improvements in the F1 score (0.797-0.949; P <.00001) and the area under the receiver operating characteristic curve (0.959-0.978; P <.00001). Four independent cohorts (3 internal and 1 external) of CRC lymph nodes were used for validation in cascading segmentation and metastasis detection models. Our approach showed excellent performance, with high sensitivity (0.995, 1.0) and specificity (0.967, 1.0) in 2 validation cohorts of adenocarcinoma cases (n = 3836 slides) when comparing slide-level labels with the ground truth (pathologist reports). Similarly, an acceptable performance was achieved in a validation cohort (n = 172 slides) with mucinous and signet-ring cell histology (sensitivity, 0.872; specificity, 0.936). The patch-based classification confidence was aggregated to overlay the potential metastatic regions within each lymph node slide for visualization. We also applied our method to a consecutive case series of lymph nodes obtained over the past 6 months at our institution (n = 217 slides). The overlays of prediction within lymph node regions matched 100% when compared with a microscope evaluation by an expert pathologist. Our results provide the basis for a computer-assisted diagnostic tool for easy and efficient lymph node screening in patients with CRC.

Deep learning: shaping the medicine of tomorrow

Predicting response to therapy is a major challenge in medicine. Machine learning algorithms are promising tools for assisting this aim. Amongst them, Deep Neural Networks are emerging as the most capable of interrogating across multiple data types. Their further development will lead to sophisticated knowledge extraction, shaping the medicine of tomorrow.

Microporous scaffolds loaded with immunomodulatory lentivirus to study the contribution of immune cell populations to tumor cell recruitment in vivo

Metastases are preceded by stochastic formation of a hospitable microenvironment known as the premetastatic niche, which has been difficult to study. Herein, we employ implantable polycaprolactone scaffolds as an engineered premetastatic niche to independently investigate the role of interleukin-10 (IL10), CXCL12, and CCL2 in recruiting immune and tumor cells and impacting breast cancer cell phenotype via lentiviral overexpression. Lentivirus delivered from scaffolds in vivo achieved sustained transgene expression for 56 days. IL10 lentiviral expression, but not CXCL12 or CCL2, significantly decreased tumor cell recruitment to scaffolds in vivo. Delivery of CXCL12 enhanced CD45+ immune cell recruitment to scaffolds while delivery of IL10 reduced immune cell recruitment. CCL2 did not alter immune cell recruitment. Tumor cell phenotype was investigated using conditioned media from immunomodulated scaffolds, with CXCL12 microenvironments reducing proliferation, and IL10 microenvironments enhancing proliferation. Migration was enhanced with CCL2 and reduced with IL10-driven microenvironments. Multiple linear regression identified populations of immune cells associated with tumor cell abundance. CD45+ immune and CD8+ T cells were associated with reduced tumor cell abundance, while CD11b+Gr1+ neutrophils and CD4+ T cells were associated with enhanced tumor cell abundance. Collectively, biomaterial scaffolds provide a tool to probe the formation and function of the premetastatic niche.

Metastasis detection from whole slide images using local features and random forests

Digital pathology has led to a demand for automated detection of regions of interest, such as cancerous tissue, from scanned whole slide images. With accurate methods using image analysis and machine learning, significant speed-up, and savings in costs through increased throughput in histological assessment could be achieved. This article describes a machine learning approach for detection of cancerous tissue from scanned whole slide images. Our method is based on feature engineering and supervised learning with a random forest model. The features extracted from the whole slide images include several local descriptors related to image texture, spatial structure, and distribution of nuclei. The method was evaluated in breast cancer metastasis detection from lymph node samples. Our results show that the method detects metastatic areas with high accuracy (AUC = 0.97-0.98 for tumor detection within whole image area, AUC = 0.84-0.91 for tumor vs. normal tissue detection) and that the method generalizes well for images from more than one laboratory. Further, the method outputs an interpretable classification model, enabling the linking of individual features to differences between tissue types. © 2017 International Society for Advancement of Cytometry.

Single-Step Assembly of Multimodal Imaging Nanocarriers: MRI and Long-Wavelength Fluorescence Imaging

Magnetic resonance imaging (MRI)- and near-infrared (NIR)-active, multimodal composite nanocarriers (CNCs) are prepared using a simple one-step process, flash nanoprecipitation (FNP). The FNP process allows for the independent control of the hydrodynamic diameter, co-core excipient and NIR dye loading, and iron oxide-based nanocrystal (IONC) content of the CNCs. In the controlled precipitation process, 10 nm IONCs are encapsulated into poly(ethylene glycol) (PEG) stabilized CNCs to make biocompatible T2 contrast agents. By adjusting the formulation, CNC size is tuned between 80 and 360 nm. Holding the CNC size constant at an intensity weighted average diameter of 99 ± 3 nm (PDI width 28 nm), the particle relaxivity varies linearly with encapsulated IONC content ranging from 66 to 533 × 10(-3) m(-1) s(-1) for CNCs formulated with 4-16 wt% IONC. To demonstrate the use of CNCs as in vivo MRI contrast agents, CNCs are surface functionalized with liver-targeting hydroxyl groups. The CNCs enable the detection of 0.8 mm(3) non-small cell lung cancer metastases in mice livers via MRI. Incorporating the hydrophobic, NIR dye tris-(porphyrinato)zinc(II) into CNCs enables complementary visualization with long-wavelength fluorescence at 800 nm. In vivo imaging demonstrates the ability of CNCs to act both as MRI and fluorescent imaging agents.

Is recombinant human TSH as effective as thyroid hormone withdrawal in the detection and treatment of metastatic differentiated thyroid cancer?

Evaluation of: Van Nostr D, Khorjekav GR, O'Neil J et al. Recombinant human thyroid-stimulating hormone versus thyroid hormone withdrawal in the identification of metastasis in differentiated thyroid cancer with ¹³¹I planar whole body imaging and ¹²⁴I PET. J. Nucl. Med 53(3), 359-362 (2012). In a recent article, Van Nostrand et al. compared two methods of patient preparation, that is, recombinant human thyroid-stimulating hormone and thyroid hormone withdrawal, for the detection of metastasis in patients previously treated for differentiated thyroid cancer, using both ¹³¹I whole-body imaging and ¹²⁴I positron emission tomography. The major finding of this prospective study was that the number of patients and the number of foci for patient positive at ¹³¹I whole-body imaging and ¹²⁴I PET performed after thyroid hormone withdrawal was significantly higher when compared with that related to patients prepared with recombinant human thyroid-stimulating hormone.