Multi-omic machine learning predictor of breast cancer therapy response.

Breast cancers are complex ecosystems of malignant cells and the tumour microenvironment1. The composition of these tumour ecosystems and interactions within them contribute to responses to cytotoxic therapy2. Efforts to build response predictors have not incorporated this knowledge. We collected clinical, digital pathology, genomic and transcriptomic profiles of pre-treatment biopsies of breast tumours from 168 patients treated with chemotherapy with or without HER2 (encoded by ERBB2)-targeted therapy before surgery. Pathology end points (complete response or residual disease) at surgery3 were then correlated with multi-omic features in these diagnostic biopsies. Here we show that response to treatment is modulated by the pre-treated tumour ecosystem, and its multi-omics landscape can be integrated in predictive models using machine learning. The degree of residual disease following therapy is monotonically associated with pre-therapy features, including tumour mutational and copy number landscapes, tumour proliferation, immune infiltration and T cell dysfunction and exclusion. Combining these features into a multi-omic machine learning model predicted a pathological complete response in an external validation cohort (75 patients) with an area under the curve of 0.87. In conclusion, response to therapy is determined by the baseline characteristics of the totality of the tumour ecosystem captured through data integration and machine learning. This approach could be used to develop predictors for other cancers.

The mechanism of balloon Eustachian tuboplasty: a biomechanical study.

Obstructive Eustachian tube dysfunction (OETD) is a common condition resulting from inadequate opening of the Eustachian tube (ET). A new surgical treatment involves high-pressure inflation of a balloon within the ET, with the aim of dilating the soft tissue structure. However, the mechanical effects of this intervention have not been established, nor the impact of changing device size or other technical parameters. A novel experimental technique allowed quantification of plastic and elastic tissue deformation in model materials and then human cadaver ETs during balloon dilation, based on the measured balloon inflation pressure-volume relationship. Plastic tissue deformation was found to be greater using larger balloons and deeper device insertion, but increasing the inflation pressure had a more limited effect, with most deformation occurring well below the clinically used pressures. Histological assessment of ET tissue suggested that mucosal tearing and cartilage cracking were in part responsible for the mechanical changes. Balloon dilation of the ET has huge potential if found to be clinically effective, but currently there is a need to understand and develop the technique further. The novel methods employed in this study will be valuable in future laboratory and in vivo studies of ET balloon dilation. Pressures are reported in Bar as this unit is used for medical balloon dilation procedures in clinical practice. 1 Bar = 100,000 Pa. Graphical abstract caption Dilation of the Eustachian tube for obstructive dysfunction is performed clinically with 3- and 6-mm-diameter balloons of approximately the same overall length. Our data suggest that dilation with a 6-mm balloon causes greater deformation of the soft tissue structure than dilation with a 3-mm balloon. This difference has yet to be demonstrated clinically. Plastic deformation was measured in terms of energy (J) dissipated during balloon inflation.

The Genomic and Immune Landscapes of Lethal Metastatic Breast Cancer.

The detailed molecular characterization of lethal cancers is a prerequisite to understanding resistance to therapy and escape from cancer immunoediting. We performed extensive multi-platform profiling of multi-regional metastases in autopsies from 10 patients with therapy-resistant breast cancer. The integrated genomic and immune landscapes show that metastases propagate and evolve as communities of clones, reveal their predicted neo-antigen landscapes, and show that they can accumulate HLA loss of heterozygosity (LOH). The data further identify variable tumor microenvironments and reveal, through analyses of T cell receptor repertoires, that adaptive immune responses appear to co-evolve with the metastatic genomes. These findings reveal in fine detail the landscapes of lethal metastatic breast cancer.

miR-146a and miR-155 Expression Levels in Acute Graft-Versus-Host Disease Incidence.

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a curative treatment for numerous hematological malignancies. However, acute graft-versus-host disease (aGVHD) is still the major complication causing mortality. MicroRNAs (miRNAs) play a significant role in inflammation and have potential as prognostic and diagnostic biomarkers. This study investigated the role of two immune-specific miRNAs (miR-146a and miR-155) as biomarkers for aGVHD incidence in the peripheral blood of allo-HSCT patients prior to disease onset. The study showed that miR-146a and its statistical interaction with miR-155 at day +28 were predictive of aGVHD incidence. Interestingly, the expression levels of miR-146a and miR-155 negatively correlated with the transcription factor, SPI1 (PU.1gene) mRNA expression.

A case of necrotising fasciitis caused by Serratia marsescens: extreme age as functional immunosuppression?

We report the case of a 97-year-old woman who had a prolonged hospital admission for the treatment of right-sided heart failure. During her stay she experienced a rapid deterioration, characterised by shortness of breath, cardiovascular compromise and a hot, red, swollen calf. Post-mortem examination demonstrated that this was caused by necrotising fasciitis due to Serratia marcescens as a single pathogen. This is only the second reported case of this condition in the absence of diabetes or immunosuppression, and clinical deterioration was much more rapid. The case underlines the importance of circumspection and regular review in the diagnosis of the elderly patient. It reminds us that these patients should be viewed as functionally immunosuppressed, and that some or all of the haematological markers of infection can be absent even in severe disease.

Auditory temporal-envelope processing in high-functioning children with Autism Spectrum Disorder.

Individuals with Autism Spectrum Disorder (ASD) perform worse than controls when listening to speech in a temporally modulated noise (Alcántara, Weisblatt, Moore, & Bolton, 2004; Groen et al., 2009). The current study examined whether this is due to poor auditory temporal-envelope processing. Temporal modulation transfer functions were measured in 6 high-functioning children with ASD and 6 control listeners, using sinusoidal amplitude modulation of a broadband noise. Modulation-depth thresholds at low modulation rates were significantly higher for the ASD group than for the Control group, and generally higher at all modulation rates tested. Low-pass filter model estimates of temporal-envelope resolution and temporal-processing efficiency showed significant differences between the groups for modulation-depth threshold values at low modulation rates. Intensity increment-detection thresholds, measured on a subset of individuals in the ASD and Control groups, were not significantly different. The results are consistent with ASD individuals having reduced processing efficiency of temporal modulations. Possible neural mechanisms that might underlie these findings are discussed.

Effects of level and frequency on the audibility of partials in inharmonic complex tones.

The effect of level and frequency on the audibility of partials was measured for complex tones with partials uniformly spaced on an equivalent rectangular bandwidth (ERB(N)) number scale. On each trial, subjects heard a sinusoidal "probe" followed by a complex tone. The probe was mistuned downwards or upwards (at random) by 4.5% from the frequency of one randomly selected partial in the complex. The subject indicated whether the probe was higher or lower in frequency than the nearest partial in the complex. The frequencies were roved from trial to trial, keeping frequency ratios fixed. In experiment 1, the level per partial, L, was 40 or 70 dB SPL and the mean frequency of the central partial, f(c), was 1201 Hz. Scores for the highest and lowest partials in the complexes were generally high for all spacings. Scores for the inner partials were close to chance at 0.75-ERB(N) spacing, and improved as the spacing was increased up to 2 ERB(N). For intermediate spacings, performance was better for the lower level used. In experiment 2, L was 70 dB SPL and f(c) was 3544 Hz. Performance worsened markedly for partial frequencies above 3544 Hz, consistent with a role of phase locking.