This article presents Individual Conditional Expectation (ICE) plots, a tool for visualizing the model estimated by any supervised learning algorithm. Classical partial dependence plots (PDPs) help visualize the average partial relationship between the predicted response and one or more features. In the presence of substantial interaction effects, the partial response relationship can be heterogeneous. Thus, an average curve, such as the PDP, can obfuscate the complexity of the modeled relationship. Accordingly, ICE plots refine the partial dependence plot by graphing the functional relationship between the predicted response and the feature for individual observations. Specifically, ICE plots highlight the variation in the fitted values across the range of a covariate, suggesting where and to what extent heterogeneities might exist. In addition to providing a plotting suite for exploratory analysis, we include a visual test for additive structure in the data generating model. Through simulated examples and real data sets, we demonstrate how ICE plots can shed light on estimated models in ways PDPs cannot. Procedures outlined are available in the R package ICEbox.

In the early 1980s, Halbert White inaugurated a "model-robust" form of statistical inference based on the "sandwich estimator" of standard error. This estimator is known to be "heteroskedasticity-consistent," but it is less well known to be "nonlinearity-consistent" as well. Nonlinearity, however, raises fundamental issues because in its presence regressors are not ancillary, hence cannot be treated as fixed. The consequences are deep: (1) population slopes need to be reinterpreted as statistical functionals obtained from OLS fits to largely arbitrary joint x-y distributions; (2) the meaning of slope parameters needs to be rethought; (3) the regressor distribution affects the slope parameters; (4) randomness of the regressors becomes a source of sampling variability in slope estimates of order 1/ √ N;(5) inference needs to be based on model-robust standard errors, including sandwich estimators or the x-y bootstrap. In theory, model-robust and model-trusting standard errors can deviate by arbitrary magnitudes either way. In practice, significant deviations between them can be detected with a diagnostic test.

The stress-inducible 72 kDa heat shock protein Hsp70 is known to be expressed on the membrane of highly aggressive tumor cells including high-grade gliomas, but not on the corresponding normal cells. Membrane Hsp70 (mHsp70) is rapidly internalized into tumor cells and thus targeting of mHsp70 might provide a promising strategy for theranostics. Superparamagnetic iron oxide nanoparticles (SPIONs) are contrast negative agents that are used for the detection of tumors with MRI. Herein, we conjugated the Hsp70-specific antibody (cmHsp70.1) which is known to recognize mHsp70 to superparamagnetic iron nanoparticles to assess tumor-specific targeting before and after ionizing irradiation. In vitro experiments demonstrated the selectivity of SPION-cmHsp70.1 conjugates to free and mHsp70 in different tumor cell types (C6 glioblastoma, K562 leukemia, HeLa cervix carcinoma) in a dose-dependent manner. High-resolution MRI (11 T) on T(2)-weighted images showed the retention of the conjugates in the C6 glioma model. Accumulation of SPION-cmHsp70.1 nanoparticles in the glioma resulted in a nearly 2-fold drop of T*(2) values in comparison to non-conjugated SPIONs. Biodistribution analysis using NLR-M(2) measurements showed a 7-fold increase in the tumor-to-background (normal brain) uptake ratio of SPION-cmHsp70.1 conjugates in glioma-bearing rats in comparison to SPIONs. This accumulation within Hsp70-positive glioma was further enhanced after a single dose (10 Gy) of ionizing radiation. Elevated accumulation of the magnetic conjugates in the tumor due to radiosensitization proves the combination of radiotherapy and application of Hsp70-targeted agents in brain tumors.

Heat shock protein 70 (Hsp70) which is expressed on the plasma membrane of highly aggressive tumors including non-small cell lung carcinoma and glioblastoma multiforme serves as a target for Hsp70-targeting NK cells. Herein, we aimed to investigate the antitumor effects of a combined therapy consisting of ex vivo Hsp70-peptide TKD/IL-2-activated NK cells in combination with mouse/human anti-PD-1 antibody in a syngeneic glioblastoma and a xenograft lung cancer mouse model. Mice with membrane Hsp70 positive syngeneic GL261 glioblastoma or human xenograft A549 lung tumors were sham-treated with PBS or injected with ex vivo TKD/IL-2-activated mouse/human NK cells and mouse/human PD-1 antibody either as a single regimen or in combination. Tumor volume was assessed by MR scanning and tumor-infiltrating CD8 + T, NK, and PD-1 + cells were quantified by immunohistochemistry (IHC). We could show that the adoptive transfer of ex vivo TKD/IL-2-activated mouse NK cells or the inhibition of PD-1 resulted in tumor growth delay and an improved overall survival (OS) in a syngeneic glioblastoma mouse model. A combination of both therapies was well-tolerated and significantly more effective with respect to both outcome parameters than either of the single regimens. A combined treatment in a xenograft lung cancer model showed identical effects in immunodeficient mice bearing human lung cancer after adoptive transfer of TKD/IL-2-activated human effector cells and a human PD-1 antibody. Tumor control was associated with a massive infiltration with CD8 + T and NK cells in both tumor models and a decreased in PD-1 expression on immune effector cells. In summary, a combined approach consisting of activated NK cells and anti-PD-1 therapy is safe and results in a long-term tumor control which is accompanied by a massive tumor immune cell infiltration in 2 preclinical tumor models.

Chaperone Hsp70 can activate adaptive immunity suggesting its possible application as an antitumor vaccine. To assess the therapeutic capacity of Hsp70 we administered purified chaperone into a C6 glioblastoma brain tumor and explored the viability and tumor size as well as interferon gamma (IFNc) production and cytotoxicity of lymphocytes in the treated animals. Targeted intratumoral injection of Hsp70 resulted in its distribution within the area of glioblastoma, and caused significant inhibition of tumor progression as confirmed by magnetic resonance imaging. The delay in tumor growth corresponded to the prolonged survival of tumor-bearing animals of up to 31 days versus 20 days in control. Continuous administration of Hsp70 with an osmotic pump increased survival even further (39 days). Therapeutic efficacy was associated with infiltration to glioblastoma of NK cells (Ly-6c1) and T lymphocytes (CD31, CD41 and CD81) as well as with an increase in the activity of NK cells (granzyme B production) and CD81 T lymphocytes as shown by IFNc ELISPOT assay. Furthermore, we found that Hsp70 treatment caused concomitantly, with a tenfold elevated IFNc production, an increase in anti-C6 tumor cytotoxicity of lymphocytes. In conclusion, continuous intratumoral delivery of Hsp70 demonstrates high therapeutic potential and therefore could be applied in the treatment of glioblastoma.Multiform glioblastoma is the most aggressive and common primary brain tumor in adults 1 with an incidence rate of 7.3 cases per 100,000 persons per year.2 The standard treatment for newly diagnosed glioblastoma currently includes surgical resection followed by radiotherapy and chemotherapy using apoptosis-inducing factors such as temozolomide, which is actively used nowadays. 3During the last two decades, increasing evidence has demonstrated that the high immunomodulatory potency of heat shock protein 70 (Hsp70) is due to its ability to specifically bind tumor antigens and present them to dendritic antigenpresenting cells (APCs).4 Several anticancer and antivirus vaccines were created based on Hsp70 chaperone. First, immunization of mice bearing tumors with Hsp70-secreting murine ovarian cancer cells caused a strong anticancer effect concomitantly with the elicitation of CD41 and CD81 Tand NK-cell activity.5 Another chaperonic vaccine contains fusion products of Hsp70 with Herpes virus VP22 peptide (aa268-301) that facilitate intracellular transport; this construct was effective in inducing a specific cytotoxic response against mouse lymphoma cells.6 Third, Ito et al. showed that the combined intratumoral delivery of pure Hsp70 with heating magnetic particles can efficiently destroy mouse melanoma B16 in a therapeutic modality. 7 More recently, we reported that the application of purified Hsp70 onto skin area covering B16/F10 melanoma significantly inhibited tumor progression and increased the survival of animals. 8Although the mechanism of antitumor activity of Hsp70 remains unclear, it is believed that the chaperone can trigger both innate and ...