 Research
 Open Access
 Published:
Leveraging auxiliary measures: a deep multitask neural network for predictive modeling in clinical research
BMC Medical Informatics and Decision Makingvolume 18, Article number: 126 (2018)
Abstract
Background
Accurate predictive modeling in clinical research enables effective early intervention that patients are most likely to benefit from. However, due to the complex biological nature of disease progression, capturing the highly nonlinear information from lowlevel input features is quite challenging. This requires predictive models with highcapacity. In practice, clinical datasets are often of limited size, bringing danger of overfitting for highcapacity models. To address these two challenges, we propose a deep multitask neural network for predictive modeling.
Methods
The proposed network leverages clinical measures as auxiliary targets that are related to the primary target. The predictions for the primary and auxiliary targets are made simultaneously by the neural network. Network structure is specifically designed to capture the clinical relevance by learning a shared feature representation between the primary and auxiliary targets. We apply the proposed model in a hypertension dataset and a breast cancer dataset, where the primary tasks are to predict the left ventricular mass indexed to body surface area and the time of recurrence of breast cancer. Moreover, we analyze the weights of the proposed neural network to rank input features for model interpretability.
Results
The experimental results indicate that the proposed model outperforms other different models, achieving the best predictive accuracy (mean squared error 199.76 for hypertension data, 860.62 for Wisconsin prognostic breast cancer data) with the ability to rank features according to their contributions to the targets. The ranking is supported by previous related research.
Conclusion
We propose a novel effective method for clinical predictive modeling by combing the deep neural network and multitask learning. By leveraging auxiliary measures clinically related to the primary target, our method improves the predictive accuracy. Based on featue ranking, our model is interpreted and shows consistency with previous studies on cardiovascular diseases and cancers.
Background
Accurate prediction for disease phenotypes is one of the most important tasks in clinical research, as it can enable effective early interventions that patients are most likely to benefit. Due to the intrinsic complex biological mechanism of disease progression, successful predictive models should be capable of learning highlevel information from lowlevel input features. However, traditional methods, such as linear regression, simplify the disease progression as additive effects of input features (i.e. age, blood pressure, renal function). Consequently, nonadditive relations are not captured, potentially leading to less satisfactory predictive performances.
Deep neural networks (DNNs) have achieved great improvements for difficult predictive tasks in speech recognition, computer vision and healthcare informatics [1–4]. Compared with linear regression, DNNs have the capability of learning highlevel feature representations, rendering better predictions based on those abstract features. This enables DNN to capture the nonlinear relations of lowlevel features, making itself promising in clinical research.
Successful deep neural networks require abundant labeled data for effectively learning useful feature representations. However in clinical practice, collecting labeled data is expensive and timeconsuming. As a result, only a limited amount of labeled data are available. Fitting a highcapacity model could potentially overfit the small amount of labeled data.
To avoid overfitting of DNNs, various regularization methods, such as dropout, early stopping and L2 regularization [5] have been developed. In the domain of clinic research, with defining primary targets, we can further mitigate overfitting by leveraging other clinical measures that are generated by the labeling process. As these measures are clinically related to the primary targets, we can integrate them into multitask framework as regularization that can benefit our model.
For instance, some demographic subpopulations with hypertension are more likely to develop left ventricular hypertrophy (LVH), a form of structural heart damage that results from poor blood pressure control. Left ventricular mass indexed to body surface area (LVMI) is a commonly used method of determining when LVH is present. However, measuring LVMI requires advanced imaging but it is difficult to know which patients should undergo testing, it is challenging to predict as there is no single input features having enough explanation power for LVH. Accurately predicting LVH status for hypertension patients is critical as definitive testing to diagnose LVH, including cardiac magnetic resonance imaging (CMR), is expensive and testing every patient with hypertension would be cost prohibitive.
In this paper, we propose generalized auxiliarytask augmented network (GATAN), extending [6] from regression to general supervised learning tasks. GATAN is a multitask predictive neural network that predicts the primary target and auxiliary targets simultaneously (See Fig. 1). Under the multitask learning framework, the auxiliary tasks can be viewed as a regularization method as well as implicit data augmentation [5, 7–10]. GATAN hence can reduce the risk of overfitting. Without a universal definition of intertask relatedness, GATAN learns taskspecific feature representations, as well as a shared representation for all tasks to conceptually capture the relation. The learned representations are then combined together using a weighting mechanism; GATAN makes predictions based on the combined highlevel features. Finally, to interpret GATAN, we adopt a heuristic method that analyze the learned weights to rank the contribution of input features.
Methods
Generalzied auxiliarytask augmented network
Taking the motivating example in Fig. 1, with LVMI being the primary target, the labeling process would produce additional CMR measures that are also characteristics of heart morphology including septal, posterior and anterior heart wall thickness. These measures are clinically related to LVMI and predictive models can exploit them as auxiliary predictive tasks.
However,the clinical “relevance” is not clearly defined. To circumvent this issue, GATAN models learns a feature representation that can be decomposed into a weighted sum of the shared and taskspecific feature representation. The shared representation conceptually models the relevance between tasks. Figure 2 displays GATAN structure. We use feedforward deep neural network (FDNN) [5] as the building block for GATAN.
Assume that\(\left ({\vphantom {y^{c}_{a}}}\mathbf {x}, y^{c}, y^{a}\right)\) is a sample with input features x, primary target y^{c} and auxiliary target y^{a}. The shared and taskspecific feature representations are learned as follows:
where f(·) is modeled by FDNN with multiple stacked hidden layers and nonlinear activation (elementwise sigmoid action in our case). These feature representations are then combined to form the final representations h^{fc} and h^{fa}:
where {a_{1},a_{2}} and {b_{1},b_{2}} are the weights that quantify the contributions of h^{s}, h^{c} and h^{a}. Note that in this formulation, h^{s}, h^{c} and h^{a} are of the same dimension. As a side note, another strategy to combine the taskspecific and shared feature representations is through vector concatenation h^{fc}=[h^{c},h^{s}]. But this approach could introduce more parameters for each h having enough representation power. We hence prefer the weighted sum approach when only limited amount of data is available.
To compute {a_{1},a_{2}} and {b_{1},b_{2}}, the cosinedistance “cosd” is used:
for the primary task, and
for the auxiliary task, where cosd (v_{1},v_{2})=v_{1}·v_{2}/(v_{1}_{2}v_{2}_{2}), ·_{2} is the euclidean norm of a vector. Since we use sigmoid as the activation function, {a_{1},a_{2}} and {b_{1},b_{2}} are positive and hence proper weights. Note that this strategy biases toward the shared feature representation and forces it to makes at least half contribution (i.e. a_{2},b_{2}≥0.5) to the final feature representation for GATAN, displaying the benefits of multitask learning.
Based on the final feature representation, the prediction \(\hat {y}^{c}\) and \(\hat {y}^{a}\) are calculated:
where W^{c} and W^{a} are dimensioncompatible vectors, h^{c} and h^{a} are bias terms, and l(·) is the link function depending on the specific prediction tasks. When targets are continuous, l(·) is the identity function; for classification tasks, l(·) is the sigmoid or softmax function:
The joint objective function is a sum of the loss function for each task:
where for notational brevity, we use Θ to represent the set of parameters in the neural network, ω is a hyperparameter balancing different tasks during training. We use ω=1 in our experiments.
For regression, the loss function is the squared loss:
For classification, it is crossentropy:
where we have encoded y^{c} as onehot vector.
Note that GATAN also allows multiple auxiliary targets which can be incorporated into GATAN straightforwardly, and different types of loss functions for different tasks such as one regression task and one classification task.
Feature ranking
Model interpretability is another important aspect in clinical practice. While there are no systematic ways to interpret deep networks, we can extend from linear regression to calculate the contribution of each input feature by back propagating each neuron’s contribution through its connections to previous layer of neurons [11].
To see the backpropagation of each neuron’s contribution to the target, let us take an example shown in Fig. 3. Let \(\mathbf {W_{1}} = \left (w_{ij}^{1}\right)_{3\times 2}\) and \(\mathbf {W_{2}} = \left (w_{ij}^{2}\right)_{1\times 3}\) be the two weight matrices associated with the last two hidden layers.
h_{j}’s contribution can be computed as in linear regression for {j:1,2,3}:
Similarly, g_{k}’s (k=1,2) contribution C_{kj} to h_{j} is
Then the contribution C_{kjy} from g_{k}through h_{j} to y is
Since there are three paths from g_{k} to y through h_{1}, h_{2} and h_{3}, the total contribution C_{ky} of g_{k} is
We can keep propagating the contribution of neurons to input features to calculate their contributions to the target.
In GATAN, each input features can contribute to the target through the taskspecific and the shared network. If \(C_{ky^{c}}^{c}\) and \(C_{ky^{c}}^{s}\) are the contributions of feature x_{k} through taskspecific and shared network to y^{c} respectively, the overall contribution \(\phantom {\dot {i}\!}C_{ky^{c}}\) for x_{k} is just the weighted sum given by
which provides us a heuristic approach for interpreting GATAN, a_{1} and a_{2} are given by (4).
Datasets and preprocessing
Hypertension dataset The cohort was derived from an NIHfunded study of African American patients with hypertension and elevated systolic blood pressure (>160 mm Hg) at the emergency department of Detroit Receiving Hospital. Previous studies have shown that there are disparities among hypertension patients with some who are at greater risk of LVH. This makes a DNN model that is capable of capturing complex feature interactions promising for predicting LVMI.
In the labeling process of LVMI, other measures that characterize heart morphology such as left ventricular stroke volume to body surface area (LVSVI), left ventricular enddiastolic volume indexed to body surface area (LVEDVI) and septal, posterior and anterior wall thickness, are also produced. These measures are closely relevant with LVMI and provides additional information that can be utilized in GATAN as auxiliary tasks.
The original dataset contains 155 samples and 65 measures. These measures consists of LVMI, 59 input features (demographics, lab results, heart functioning et al.) and 5 other CMR measures as candidates of auxiliary targets. Table 1 and Fig. 4 left panel present basic statistics of targets.
From the perspective of predictive modeling, a model only using lab results and demographics as features (34 in total) is more preferable, as they are more widely accessible and informative for disease progression, compared with the full set of features that contains heart functioning measures. Hence, we also conduct experiments with this set of features.
Wisconsin prognostic breast cancer dataset (WPBC) is a publicly available dataset in UCI repository [12]. The dataset contains 194 records of “time to recur” for breast cancer patients (after removing 4 cases of missing target values) and 32 features including tumor size, lymph node status and 30 measures computed from a digitized image of a fine needle aspirate (FNA) of a breast mass. These derived features include the mean value, standard error and largest/worst value for 10 features: radius, texture, perimeter, area, smoothness, compactness, concavity, concave points, symmetry and fractal dimension. The primary target is the “time to recurrence of breast cancer”; the auxiliary target is the recurrence state of being “recur” or “nonrecur”.
Implementations and evaluation metrics
First, various models in scikitlearn [13] are implemented for comparison, including nonparametric models (knearest neighbors (KNN), random forest (RF)), support vector regression (SVR), regularized linear regression based models (Ridge, Lasso and the multitask Lasso (MTLasso)). A 4layer perceptron (MLP4) is also implemented whose hidden layer size is matched with GATAN.
We use Pytorch [14] for building GATAN. In our experiments, each time only one CMR measure is selected as the auxiliary target. LVEDVI is for GATAN1 and posterior wall thickness for GATAN2. GATAN consists of 4 layers with the dimension of hidden layers being 80 and 40. Standard gradient descent is used to train our model.
The hypertension dataset is split into training, testing and training sets by 95/35/30. For WPBC dataset, we split the data by 134/30/30. For non neural network models, 3fold crossvalidation on the training set is performed for best hyperparameter settings. Model performances are finally reported on the testing set. We repeat this procedure 5 times.
To evaluate performance, the following three metrics are used:

Mean squared error (MSE) measures the predictive error without considering the magnitude of target:
$$\text{MSE} = \frac{1}{n}\left\left\mathbf{y}^{c}  \hat{\mathbf{y}}^{c}\right\right^{2}.$$ 
Explained variance score (EVS):
$$\text{EVS} = 1\frac{\text{Var}(\mathbf{y}^{c}  \hat{\mathbf{y}}^{c})}{\text{Var}\left(\mathbf{y}^{c}\right)},$$where Var (·) is the variance.

Median absolute error (MAE) is a more robust error than MSE that compute the median of absolute predictive errors:
$$\text{MAE} = \text{Median}\left(\left\mathbf{y}^{c}  \hat{\mathbf{y}}^{c}\right\right).$$
Smaller MSE and MAE are better while for EVS, larger is better.
Results and discussions
Hypertension data
Using entire feature set We first experiment with the full feature set. The predictive performance on the test data is shown in Table 2. In the table, GATAN1 and MTLasso use LVMI and LVEDVI; GATAN2 uses posterior wall thickness as the auxiliary target.
From the table, GATAN with LVEDVI as the auxiliary target (i.e. GATAN1) achieves the best predictive performance. For example, GATAN1 improves MSE approximately 3% compared with Lasso; compared with MTLasso, they also performs better with margins 5% (MSE), 13% (EVS), 2% (MAE). We can also see from the table that GATAN provides performance improvements over MLP4, due to the introduction of auxiliary tasks. This confirms that GATAN benefits from the auxiliary task in multitask learning as a regularization.
MTLasso also introduces auxiliary tasks. However, MTLasso does not improve over Lasso. MTLasso assumes all tasks share the same subset of effective features. This is too restrictive for LVMI and LVEDVI having the same feature structure. On the contrary, GATAN has less restrictive assumption on defining the clinical “relevance”; GATAN captures the relevance by learning a shared feature representation. This implies that a proper assumption on the task relatedness is crucial for multitask learning.
Finally, the explained variance score (EVS) is not satisfactory for all models on the testing data. From the definition, EVS is very sensitive to poor predictions. This means that all models fail for some test samples. From the histogram of LVMI (Fig. 4), we see that data might be generated from a multimodal distribution and all models fail to capture the local data structure.
We further explored the predictive behavior of GATAN and find that models often make poor predictions at the tails of sample distribution (results no shown). For the used hypertension dataset, we find that the Pearson correlation between LVMI and calcium level is 0.79 at the right tail (LVMI >120). A twotail correlation test shows the Pearson correlation is statistically significant (pvalue <0.001). However, the Pearson correlation between LVMI and calcium is 0.00 for the entire dataset, 0.10 for LVMI <120. In previous studies [15, 16], it was shown that patients with LVH have strong positive correlation with serum calcium level compared to those without LVH. Our observations are consistent with these findings. This disparity of correlation between LVMI and calcium among the hypertension patients implies LVH prevalence differs among patient subgroups.
Using demographics and lab results only We use the same experiment setup as in the experiment with a full set of features. Table 2 shows the predictive performance with a more limited dataset. Our multitask neural network (GATAN1 and GATAN2) performs better than other models, implying that our strategy of learning highlevel feature representations would benefit predictive modeling. However, comparing with the setup of a full feature set, excluding heart functioning measures from the input features degrades model performances, as functional measures are expected to be more informative for predicting LVMI.
Interpreting GATAN Figure 5a and b show the top20 features from the full set of features with respect to two different auxiliary tasks. Comparing these two figures, we see that the feature ranking in a is approximately matched with that in a. Sex is the most important feature. In the hypertension dataset, the sample mean of male versus female is 95.78 v.s 85.21; the difference between female and male is statistically significant with pvalue <0.0001 for a twosample ttest. From the figure, we also see that other features with significant contributions are functional measures, such as ejection duration, LV ejection fraction and Cornell product (an electrocardiographic predictor of LVH). This is sensible since heart structure and function are inherently related.
Figure 6 displays the top15 features from demographics and lab results. Panel a and b are also approximately matched as those in Fig. 5. From the figure, both systolic and diastolic blood pressure are the most important features for predicting LVMI. The relationship between hypertension and LVH was the basic premise of our study. This is not surprising according to [17] that elevated blood pressure corresponds with high LVMI. Moreover, GATAN identifies more subtle relations between lab results and LVMI, including potassium, vitamin D, calcium, diabetes status, bun, renin et al. These topranked features accord with previous researches ([15, 18, 19]), demonstrating that feature ranking by analyzing the learned weights is a reasonable heuristic for interpreting deep neural networks.
WPBC data
Table 3 shows the performance of different models on the WPBC testing data. In terms of MSE and MAE, GATAN achieves the smallest predictive error 860.625 and 23.860 respectively. For the explained variance score (EVS), all models perform poorly. One reason accountable for this phenomenon is that the distribution of primary target “time to recur” is highly rightskewed, making it difficult for models fitting data with a long tail well. Figure 7 presents the top10 important features for predicting “time to recur”, including FNA area, radius and texture et al. This is intuitive as morphological measures are informative about the breast cancer.
Conclusions
In this paper, we propose a deep multitask neural network, GATAN, for predictive modeling in clinical research. GATAN leverages additional information in the modeling process by introducing clinical measures as auxiliary targets. As a DNN model, GATAN is capable of highlevel feature learning, as well as flexibly captures the clinical relevance between the primary and auxiliary targets. As our experiments using two different datasets show, with one auxiliary task demonstrate GATAN can achieve superior performance compared with traditional models when we only have access to a limited amount of labeled data.
Abbreviations
 DNN:

Deep neural network
 GATAN:

Generalized auxiliarytask augmented network
 HTN:

Hypertension
 LVH:

Left ventricular hypertrophy
 LVMI:

Left ventricular mass indexed to body surface area
References
 1
Cho K, Van Merriënboer B, Gulcehre C, Bahdanau D, Bougares F, Schwenk H, et al.Learning phrase representations using RNN encoderdecoder for statistical machine translation; 2014. arXiv preprint arXiv:061078.
 2
Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. In: Advances in neural information processing systems. 2012. p. 1097–105.
 3
Nezhad MZ, Zhu D, Li X, Yang K, Levy P. Safs: A deep feature selection approach for precision medicine. In: 2016 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). Shenzhen: IEEE: 2016. p. 501–6.
 4
Choi E, Schuetz A, Stewart WF, Sun J. Using recurrent neural network models for early detection of heart failure onset. J Am Med Inform Assoc. 2016; 24(2):361–70.
 5
Goodfellow I, Bengio Y, Courville A, Bengio Y. Deep learning. Cambridge: MIT press; 2016.
 6
Li X, Zhu D, Levy P. Predictive deep network with leveraging clinical measure as auxiliary task. In: 2017 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). Kansas City: IEEE: 2017. p. 786–91.
 7
Caruana R. Multitask learning. In: Learning to learn.1998. p. 95–133.
 8
Ruder S. An Overview of MultiTask Learning in Deep Neural Networks; 2017. arXiv preprint arXiv: 605098.
 9
Zhang Y, Yang Q. A Surey on MultiTask Learning. 2017. arXiv preprint arXiv: 708114.
 10
Wang L, Li Y, Zhou J, Zhu D, Ye J. Multitask Survival Analysis. 2017 IEEE International Conference on Data Mining (ICDM). New Orleans: IEEE; 2018, pp. 485–94.
 11
Gedeon TD. Data mining of inputs: analysing magnitude and functional measures. Int J Neural Syst. 1997; 8(02):209–18.
 12
Dheeru D, Karra Taniskidou E. UCI Machine Learning Repository. 2017. http://archive.ics.uci.edu/ml.
 13
Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, et al.Scikitlearn: Machine Learning in Python. J Mach Learn Res. 2011; 12:2825–30.
 14
Pytorch. http://pytorch.org. Accessed date: 30 July 2017.
 15
Helvacı A, Çopur B, Adaş M. Correlation between Left Ventricular Mass Index and Calcium Metabolism in Patients with Essential Hypertension. Balkan Med J. 2013; 30(1):85.
 16
Li J, Wu N, Li Y, Ye K, He M, Hu R. Crosssectional analysis of serum calcium levels for associations with left ventricular hypertrophy in normocalcemia individuals with type 2 diabetes. Cardiovasc Diabetol. 2015; 14(1):43.
 17
Devereux RB, Pickering TG, Harshfield GA, Kleinert HD, Denby L, Clark L, et al.Left ventricular hypertrophy in patients with hypertension: importance of blood pressure response to regularly recurring stress. Circulation. 1983; 68(3):470–6.
 18
ElGharbawy AH, Nadig VS, Kotchen JM, Grim CE, Sagar KB, Kaldunski M, et al.Arterial pressure, left ventricular mass, and aldosterone in essential hypertension. Hypertension. 2001; 37(3):845–50.
 19
Piovesan A, Molineri N, Casasso F, Emmolo I, Ugliengo G, Cesario F, et al.Left ventricular hypertrophy in primary hyperparathyroidism. Effects of successful parathyroidectomy. Clin Endocrinol. 1999; 50(3):321–8.
Acknowledgements
Authors would like to thank National Science Foundation and National Institutes of Health for funding our research.
Funding
This work was supported by the National Science Foundation under Grant No. 1637312 and 1451316. Collection of hypertension dataset was supported under grant NIH/NIMHD 5 R01 MD005849.
Availability of data and materials
The Wisconsin prognostic breast cancer dataset is publicly available in UCI repository https://archive.ics.uci.edu/ml/index.php. The hypertension dataset is not publicly available but can be requested by contacting Phillip Levy (plevy@med.wayne.edu).
About this supplement
This article has been published as part of BMC Medical Informatics and Decision Making Volume 18 Supplement 4, 2018: Selected articles from the IEEE BIBM International Conference on Bioinformatics & Biomedicine (BIBM) 2017: medical informatics and decision making. The full contents of the supplement are available online at https://bmcmedinformdecismak.biomedcentral.com/articles/supplements/volume18supplement4.
Author information
Affiliations
Contributions
DZ and PL planned this study. XL, DZ and PL proposed the method and wrote the manuscript. XL implemented the experiments. PL collected data. All authors read and approved the manuscript.
Corresponding author
Correspondence to Dongxiao Zhu.
Ethics declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
About this article
Published
DOI
Keywords
 Predictive modeling
 Deep neural network
 Auxiliary task
 Multitask learning