Novel Functional Radiomics for Prediction of Cardiac PET Avidity in Lung Cancer Radiotherapy

PURPOSE/OBJECTIVE(S)

Traditional methods of evaluating cardiotoxicity focus solely on radiation doses to the heart and do not incorporate functional imaging information. Functional imaging has great potential to improve the ability to provide early prediction for cardiotoxicity for lung cancer patients undergoing radiotherapy. FDG-based PET/CT imaging is routinely obtained as part of standard staging work up for lung cancer patients. Although FDG PET/CT scans are typically used to evaluate the tumor, imaging guidelines note that FDG PET/CT scans are an FDA-approved method to image for cardiac inflammation, and studies have noted that the PET cardiac signal can be predictive of clinical outcomes. The purpose of this work was to develop a radiomics model to predict clinical cardiac assessment of standard of care FDG PET/CT scans.

MATERIALS/METHODS

The study included 100 consecutive lung cancer patients treated with radiotherapy who underwent standard pre-treatment FDG-PET/CT staging scans. A clinician reviewed the PET/CT scans per clinical cardiac assessment guidelines and classified the cardiac uptake as: 0 = uniform diffuse, 1 = absent, 2 = heterogeneous, with event rates of 20%, 44%, and 35%, respectively. The heart was delineated and 200 novel functional radiomics features were selected to classify cardiac FDG uptake patterns. We divided the data into an 80% training set and a 20% test set to train and evaluate the classification models. Feature reduction was carried out using the Wilcoxon test (with Bonferroni adjusted p<0.05), hierarchical clustering, and Recursive Feature Elimination. Two automatic machine learning (AutoML) frameworks were used to determine classification models: a Random Forest Classifier (Tree-based Pipeline Optimization Tool, TPOT) and Linear Discriminant Analysis (AutoSklearn). 10-fold cross validation was carried out for training and the accuracy of the ability of the models to predict for clinical cardiac assessment is reported.

RESULTS

Fifty-one independent radiomics features were reduced to 3 clinically pertinent features (PET 2D Skewness, PET Grey Level Co-occurrence Matrix Correlation, and PET Median) using feature reduction techniques. The model selected by TPOT showed 89.8% predictive accuracy in the cross validation of the training set and 85% predictive accuracy on the test set. The model selected by AutoSklearn showed 89.7% predictive accuracy in the cross validation of the training set and 80% predictive accuracy on the test set.

CONCLUSION

The novelty of this work is that it is the first study to develop and evaluate functional cardiac radiomic features from standard of care FDG PET/CT scans with the data showing good predictive accuracy with clinical imaging evaluation. If validated, the current work provides automated methods to provide functional cardiac information using standard of care imaging that can be used as an imaging biomarker for early clinical toxicity prediction for lung cancer patients.

In vivo behavior of peptide-specific T cells during mucosal tolerance induction: antigen introduced through the mucosa of the conjunctiva elicits prolonged antigen-specific T cell priming followed by anergy

The mucosa of the conjunctiva is an important site of entry for environmental Ags as well as Ags emanating from the eye itself. However, very little is known about T cell recognition of Ag introduced through this important mucosal site. We have characterized the in vivo process of CD4 T cell recognition of Ag delivered via the conjunctival mucosa. Application of soluble OVA to the conjunctiva of BALB/c mice induced potent T cell tolerance. APC-presenting OVA peptide in vivo was only found in the submandibular lymph node and not in other lymph nodes, spleen, or nasal-associated lymphoid tissue. Similarly, in TCR transgenic DO11. 10 adoptive transfer mice, OVA-specific CD4+ T cell clonal expansion was only observed in the submandibular lymph node following conjunctival application of peptide. These experiments thus define a highly specific lymphatic drainage pathway from the conjunctiva. OVA-specific T cell clonal expansion peaked at day 3 following initiation of daily OVA administration and gradually declined during the 10-day treatment period, but remained elevated compared with nontreated adoptive transfer mice. During this period, the T cells expressed activation markers, and proliferated and secreted IL-2 in vitro in response to OVA stimulation. In contrast, these cells were unable to clonally expand in vivo, or proliferate in vitro following a subsequent OVA/CFA immunization. These results suggest that Ag applied to a mucosal site can be efficiently presented in a local draining lymph node, resulting in initial T cell priming and clonal expansion, followed by T cell anergy.

Interleukin-3 and granulocyte-macrophage colony-stimulating factor enhance the generation and function of dendritic cells

Dendritic cells, well-known for their potent antigen-presenting activity, are generally present at very low frequency in the spleens of naive mice. We examined the ability of mice to generate functional dendritic cells (DC) following exposure to the cytokines interleukin-3 (IL-3) and granulocyte-macrophage colony-stimulating factor (GM-CSF). Tumours secreting these cytokines provided a continuous stimulus resulting in a greatly increased number and frequency of DC in the spleen. These cells were purified by conventional DC isolation techniques and were found to exhibit many of the characteristics of DC from unmanipulated mice, including high allo-stimulatory activity in mixed lymphocyte reactions and expression of many similar cell surface markers. Using ovalbumin-peptide specific class I- and class II-restricted hybridomas containing the lacZ reporter gene, we found that these cytokine-generated DC had a greatly increased efficacy in the uptake and processing of particulate antigen. These cells appear to have retained the ability to ingest antigen that is generally associated with immature DC, but also exhibit the peptide/major histocompatibility complex (MHC)-presenting capabilities of mature DC. Development of an assay to measure the activity of a single DC revealed that these dual activities were the properties of the majority of the cytokine-generated DC. These findings indicate that exposure in vivo to the cytokines IL-3 and GM-CSF can result in the generation of large numbers of DC with increased capability of stimulating T cells. Thus, these cells may be important in vivo in the process of cross-priming and the subsequent generation of tumour-reactive cytotoxic T lymphocytes (CTL).

IL-3-mediated enhancement of particulate antigen presentation by macrophages

Mice were exposed to interleukin- (IL-) 3 in vivo by injection of tumor cells transfected with the IL-3 gene. At 10 days post tumor injection, bone marrow cells were recovered, pulsed with particulate antigen in the form of ovalbumin (Ova)-coated magnetic beads, and tested for their ability to present antigen via class I to an Ova/class I-restricted T cell hybridoma. Cells from IL-3-stimulated mice exhibited a marked increase in antigen presentation compared with cells from mice injected with control non-cytokine-secreting tumor cells. These cells were markedly more efficient at presenting particulate Ova antigen than in presenting soluble Ova. Based on adherence, radiation resistance, and surface markers, the cells presenting antigen appear to be in the macrophage cell lineage. These cells are susceptible to lysis by antigen-specific cytotoxic T lymphocytes, which may contribute to limiting the effectiveness of antitumor responses.

Expression of human prostate-specific antigen (PSA) in a mouse tumor cell line reduces tumorigenicity and elicits PSA-specific cytotoxic T lymphocytes

Human prostate-specific antigen (PSA) has a highly restricted tissue distribution. Its expression is essentially limited to the epithelial cells of the prostate gland. Moreover, it continues to be synthesized by prostate carcinoma cells. This makes PSA an attractive candidate for use as a target antigen in the immunotherapy of prostate cancer. As a first step in characterizing the specific immune response to PSA and its potential use as a tumor-rejection antigen, we have incorporated PSA into a well-established mouse tumor model. Line 1, a mouse lung carcinoma, and P815, a mouse mastocytoma, have been transfected with the cDNA for human PSA. Immunization with a PSA-expressing tumor cell line demonstrated a memory response to PSA which protected against subsequent challenge with PSA-expressing, but not wild-type, tumors. Tumor-infiltrating lymphocytes could be isolated from PSA-expressing tumors grown in naive hosts and were specifically cytotoxic against a syngeneic cell line that expressed PSA. Immunization with tumor cells resulted in the generation of primary and memory cytotoxic T lymphocytes (CTL) specific for PSA. The isolation of PSA-specific CTL clones from immunized animals further demonstrated that PSA can serve as a target antigen for antitumor CTL. The immunogenicity studies carried out in this mouse tumor model provide a rationale for the design of methods to elicit PSA-specific cell-mediated immunity in humans.

Analysis of the effect of cytokines (interleukins 2, 3, 4, and 6, granulocyte-monocyte colony-stimulating factor, and interferon-gamma) on generation of primary cytotoxic T lymphocytes against a weakly immunogenic tumor

We have investigated the role of cytokines (IL-2, IL-3, IL-4, IL-6, IFN-gamma, and GM-CSF) in the generation of primary cytotoxic T lymphocytes (CTL), within a single tumor system. The murine carcinoma line 1 was transfected with expression vectors with cDNA for these cytokines. Line 1 expresses low levels of class I MHC molecules, but can be induced with dimethyl sulfoxide or IFN-gamma to express high levels of class I. Class I low line 1 cells are not susceptible to CTL lysis, while class I high line 1 are lysed by CTL, which allows us to assay for CTL activity. To isolate primary CTL, tumor-infiltrating lymphocytes were isolated from cytokine-expressing or control tumors, growing in vivo. Most cytokines stimulated nonspecific killers, but IL-2 and IL-3 stimulated primary CTL. While IFN-gamma alone did not generate primary CTL, coexpression of IFN-gamma with IL-2 resulted in CTL generation. This is the first comparison of the effects of a series of cytokines on primary CTL development, and has important implications for vaccine development and immunotherapy.

Two populations of splenic dendritic cells detected with M342, a new monoclonal to an intracellular antigen of interdigitating dendritic cells and some B lymphocytes

A monoclonal has been isolated that labels an intracellular antigen in dendritic cells and some B cells. The M342 hamster immunoglobulin was selected because it stained cells in the periarterial sheaths of spleen, the deep cortex of lymph node, and the thymic medulla--the same regions in which one finds interdigitating cells, the presumptive in situ counterparts of isolated lymphoid dendritic cells. M342 labeled an antigen within granules of isolated dendritic cells, but only in cells that had been cultured for a day and not in fresh isolates. This extends recent findings that most freshly isolated spleen dendritic cells are located in the periphery of the white pulp nodule and may serve as precursors for the periarterial pool of interdigitating cells, the site for M342 staining in situ. By electron microscopic immunolabeling, the M342 antigen was found exclusively in a type of multivesicular body. M342 staining was not found in mononuclear phagocytes from blood and peritoneal cavity. Peritoneal B cells expressed M342+ granules, and upon appropriate stimulation splenic B cells developed reactive granules as well. We conclude that M342 is a strong marker for interdigitating cells. Its existence reveals intracellular specializations in the vacuolar system of antigen-presenting cells including subsets of dendritic cells.