Accurate Patient-Specific Machine Learning Models of Glioblastoma Invasion Using Transfer Learning

BACKGROUND AND PURPOSE

MR imaging-based modeling of tumor cell density can substantially improve targeted treatment of glioblastoma. Unfortunately, interpatient variability limits the predictive ability of many modeling approaches. We present a transfer learning method that generates individualized patient models, grounded in the wealth of population data, while also detecting and adjusting for interpatient variabilities based on each patient's own histologic data.

MATERIALS AND METHODS

We recruited patients with primary glioblastoma undergoing image-guided biopsies and preoperative imaging, including contrast-enhanced MR imaging, dynamic susceptibility contrast MR imaging, and diffusion tensor imaging. We calculated relative cerebral blood volume from DSC-MR imaging and mean diffusivity and fractional anisotropy from DTI. Following image coregistration, we assessed tumor cell density for each biopsy and identified corresponding localized MR imaging measurements. We then explored a range of univariate and multivariate predictive models of tumor cell density based on MR imaging measurements in a generalized one-model-fits-all approach. We then implemented both univariate and multivariate individualized transfer learning predictive models, which harness the available population-level data but allow individual variability in their predictions. Finally, we compared Pearson correlation coefficients and mean absolute error between the individualized transfer learning and generalized one-model-fits-all models.

RESULTS

Tumor cell density significantly correlated with relative CBV (r = 0.33, P < .001), and T1-weighted postcontrast (r = 0.36, P < .001) on univariate analysis after correcting for multiple comparisons. With single-variable modeling (using relative CBV), transfer learning increased predictive performance (r = 0.53, mean absolute error = 15.19%) compared with one-model-fits-all (r = 0.27, mean absolute error = 17.79%). With multivariate modeling, transfer learning further improved performance (r = 0.88, mean absolute error = 5.66%) compared with one-model-fits-all (r = 0.39, mean absolute error = 16.55%).

CONCLUSIONS

Transfer learning significantly improves predictive modeling performance for quantifying tumor cell density in glioblastoma.

Effect of Poultry Litter Treatment (PLT) on death due to ascites in broilers

The purposes of this study were to determine the effect of Poultry Litter Treatment (PLT) on levels of litter moisture, litter nitrogen, atmospheric ammonia, and death due to ascites. Data were collected from chicks raised in containment conditions that resembled commercial settings. The ascites death rate (5.9%) in broiler chicks on PLT-treated litter was significantly (chi 2 = 15.5, df = 1, P = 0.0001) lower than that (31.5%) in broiler chicks raised on untreated litter. Likewise, atmospheric ammonia levels in pens that had been treated with PLT were significantly (P < 0.05) lower than those in pens that received no treatment. Under the conditions of the present study, litter moisture and litter nitrogen levels were not different (P > 0.05) among treatments at any sample interval.

Kappa opioid receptors of human placental villi modulate acetylcholine release

Human placental villus tissue is non-innervated, yet it contains components of the opiate and cholinergic systems. We investigated whether opioids modulate a calcium dependent acetylcholine release from the villus tissue in a manner similar to that demonstrated by the parasympathetic nerve-smooth muscle junction. We reported that the kappa receptor agonist ethylketocyclazocine (EKC) inhibits acetylcholine release, and that the inhibition is reversed by the selective antagonist, Mr2266. Findings reported here substantiate the role of opioids as modulators of acetylcholine release from villus tissue. The nonselective agonist, morphine, also inhibits acetylcholine release. Inhibition caused by morphine is reversed by low concentrations of non-selective antagonists, naloxone and naltrexone. Naloxone at high concentrations potentiates the inhibition of acetylcholine release caused by morphine. In addition, the calcium channel blocker, diltiazem, was found to inhibit the release of acetylcholine. The combination of morphine and diltiazem resulted in a greater inhibition of acetylcholine release than by either alone. These results suggest that opiate cholinergic interactions occur in non-neural tissue with a mechanism similar to that known to occur at certain cholinergic synapses.

Efficacy of maduramicin against ionophore-tolerant field isolates of coccidia in broilers

Maduramicin ammonium was given at 2.5-8 ppm in the feed to broilers experimentally infected with coccidia recently isolated from broiler farms where ionophores had been used for several years. Infection pressure varied from mild to severe in five trials: mortality in unmedicated controls ranged from 0 to 59%, intestinal lesion scores were high, and weight gain was depressed by the infections. The cultures of Eimeria were partly resistant to ionophores: birds medicated with monensin at 100-121 ppm had only modest reductions in lesion scores and incomplete protection against weight loss or mortality. Control of infections by maduramicin was significant at 4 ppm but best at 5-7 ppm. Maduramicin was more effective than monensin or narasin, but about the same as salinomycin, in reducing lesions and mortality and in protecting performance. Maduramicin was well tolerated within the dose range of 5-7 ppm.