Combined Artificial Intelligence Approaches Analyzing 1000 Conservative Patients with Back Pain—A Methodological Pathway to Predicting Treatment Efficacy and Diagnostic Groups

Patients with back pain are common and present a challenge in everyday medical practice due to the multitude of possible causes and the individual effects of treatments. Predicting causes and therapy efficien cy with the help of artificial intelligence could improve and simplify the treatment. In an exemplary collective of 1000 conservatively treated back pain patients, it was investigated whether the prediction of therapy efficiency and the underlying diagnosis is possible by combining different artificial intelligence approaches. For this purpose, supervised and unsupervised artificial intelligence methods were analyzed and a methodology for combining the predictions was developed. Supervised AI is suitable for predicting therapy efficiency at the borderline of minimal clinical difference. Non-supervised AI can show patterns in the dataset. We can show that the identification of the underlying diagnostic groups only becomes possible through a combination of different AI approaches and the baseline data. The presented methodology for the combined application of artificial intelligence algorithms shows a transferable path to establish correlations in heterogeneous data sets when individual AI approaches only provide weak results.

Download Full-text

Chronic Pain Patients Exhibit Individually Unique Cortical Signatures of Pain

10.1101/2020.09.05.284117 ◽

2020 ◽

Author(s):

Astrid Mayr ◽

Pauline Jahn ◽

Anne Stankewitz ◽

Bettina Deak ◽

Anderson Winkler ◽

...

Keyword(s):

Chronic Pain ◽

Back Pain ◽

Pain Intensity ◽

Chronic Back Pain ◽

Anterior Insula ◽

Individual Level ◽

Intensity Changes ◽

Complex Picture ◽

The Individual ◽

Pain Patients

AbstractWe investigated how the trajectory of pain patients’ ongoing and fluctuating pain is encoded in the brain. In repeated fMRI sessions, 20 chronic back pain patients and 20 chronic migraineurs were asked to continuously rate the intensity of their endogenous pain. Linear mixed effects models were used to disentangle cortical processes related to pain intensity and to pain intensity changes. We found that the intensity of pain in chronic back pain patients is encoded in the anterior insula, the frontal operculum, and the pons; the change of pain of chronic back pain and chronic migraine patients is mainly encoded in the anterior insula. At the individual level, we identified a more complex picture where each patient exhibited their own signature of endogenous pain encoding. The diversity of the individual cortical signatures of chronic pain encoding results adds to the understanding of chronic pain as a complex and multifaceted disease.

Download Full-text

Individually Unique Dynamics of Cortical Connectivity Reflect the Ongoing Intensity of Chronic Pain

10.1101/2021.06.30.450553 ◽

2021 ◽

Author(s):

Astrid Mayr ◽

Pauline Jahn ◽

Bettina Deak ◽

Anne Stankewitz ◽

Vasudev Devulapally ◽

...

Keyword(s):

Chronic Pain ◽

Back Pain ◽

Pain Intensity ◽

Chronic Back Pain ◽

Personalised Medicine ◽

Brain Regions ◽

Cortical Connectivity ◽

Wide Range ◽

The Individual ◽

Pain Patients

Background. Chronic pain diseases are characterised by an ongoing and fluctuating endogenous pain, yet it remains to be elucidated how this is reflected by the dynamics of ongoing functional cortical connections. The present study addresses this disparity by taking the individual perspective of pain patients into account, which is the varying intensity of endogenous pain. Methods. To this end, we investigated the cortical encoding of 20 chronic back pain patients and 20 chronic migraineurs in four repeated fMRI sessions. During the recording, the patients were asked to continuously rate their pain intensity. A brain parcellation approach subdivided the whole brain into 408 regions. A 10 s sliding-window connectivity analysis computed the pair-wise and time-varying connectivity between all brain regions across the entire recording period. Linear mixed effects models were fitted for each pair of brain regions to explore the relationship between cortical connectivity and the observed trajectory of the patients' fluctuating endogenous pain. Results. Two pain processing entities were taken into account: pain intensity (high, middle, low pain) and the direction of pain intensity changes (rising vs. falling pain). Overall, we found that periods of high and increasing pain were predominantly related to low cortical connectivity. For chronic back pain this applies to the pain intensity-related connectivity for limbic and cingulate areas, and for the precuneus. The change of pain intensity was subserved by connections in left parietal opercular regions, right insular regions, as well as large parts of the parietal, cingular and motor cortices. The change of pain intensity direction in chronic migraine was reflected by decreasing connectivity between the anterior insular cortex and orbitofrontal areas, as well as between the PCC and frontal and ACC regions. Conclusions. Interestingly, the group results were not mirrored by the individual patterns of pain-related connectivity, which is suggested to deny the idea of a common neuronal core problem for chronic pain diseases. In a similar vein, our findings are supported by the experience of clinicians, who encounter patients with a unique composition of characteristics: personality traits, various combinations of symptoms, and a wide range of individual responses to treatment. The diversity of the individual cortical signatures of chronic pain encoding results adds to the understanding of chronic pain as a complex and multifaceted disease. The present findings support recent developments for more personalised medicine.

Download Full-text

The clinical back pain courses described by information available in Danish central registries

BMC Health Services Research ◽

10.1186/s12913-021-07409-w ◽

2022 ◽

Vol 22 (1) ◽

Author(s):

Maria Iachina ◽

Olav S. Garvik ◽

Pernille S. Ljungdalh ◽

Mette Wod ◽

Berit Schiøttz-Christensen

Keyword(s):

Back Pain ◽

Clinical Course ◽

Diagnosis And Treatment ◽

Time Interval ◽

Pain Diagnosis ◽

The Everyday ◽

Previous Contact ◽

One Year ◽

The Individual ◽

Pain Patients

Abstract Background Patients with back pain are often in contact with 2–4 hospital departments when receiving a back pain diagnosis and treatment. This complicates the entire clinical course description. There is, currently, no model that describes the course across departments for patients with back pain. This study aims to construct an interdisciplinary clinical course using the central register’s information. Methods All patients with back pain referred for diagnosis and treatment at the Spine Center of Southern Denmark from 1 January 2011 until 31 December 2017 were included. By means of information available in central registers, we described the interdisciplinary clinical course for the individual patient, including information on all contacts at different departments, and proposed three different models to define the index and final date. The index date was defined as the first visit without a previous contact to the Spine Center for 6 months for model I, 1 year for model II, and 2 years for model III. The final date was defined as the last visit without following contacts for 6 months, 1 year, and 2 years, respectively, for models I, II, and III. Results A total of 69,564 patients (male: n = 30,976) with back pain diagnosis were identified. The three models all leave the information on the entire course at the hospital. In model I (64,757 clinical back pain courses), the time span to a possible previous clinical course is too short to secure the start of a new course (14% had two or more). With at least 1 year between a possible previous contact, model II (60,914 courses) fits the everyday clinical practice (9% had two or more clinical back pain courses). In model III (60,173 courses) it seems that two independent courses might be connected in the same course as only 5% had two or more clinical back pain courses. Conclusions Despite contact with different departments, the clinical course for back pain patients can be described by information from the central registers. A one-year time interval fits best the clinicians’ everyday observations.

Download Full-text

INTEGRATING MICROARRAY DATA BY CONSENSUS CLUSTERING

International Journal of Artificial Intelligence Tools ◽

10.1142/s0218213004001867 ◽

2004 ◽

Vol 13 (04) ◽

pp. 863-880 ◽

Cited By ~ 64

Author(s):

VLADIMIR FILKOV ◽

STEVEN SKIENA

Keyword(s):

Microarray Data ◽

Heterogeneous Data ◽

Data Sets ◽

General Criterion ◽

Consensus Clustering ◽

Individual Data ◽

Set Partition ◽

Data Set ◽

The Individual ◽

General Method

With the exploding volume of microarray experiments comes increasing interest in mining repositories of such data. Meaningfully combining results from varied experiments on an equal basis is a challenging task. Here we propose a general method for integrating heterogeneous data sets based on the consensus clustering formalism. Our method analyzes source-specific clusterings and identifies a consensus set-partition which is as close as possible to all of them. We develop a general criterion to assess the potential benefit of integrating multiple heterogeneous data sets, i.e. whether the integrated data is more informative than the individual data sets. We apply our methods on two popular sets of microarray data yielding gene classifications of potentially greater interest than could be derived from the analysis of each individual data set.

Download Full-text

Nonspecific Chronic Back Pain: Sixth Edition Approaches

Guides Newsletter ◽

10.1001/amaguidesnewsletters.2019.sepoct03 ◽

2019 ◽

Vol 24 (5) ◽

pp. 14-15

Author(s):

Jay Blaisdell ◽

James B. Talmage

Keyword(s):

Back Pain ◽

Chronic Back Pain ◽

Intervertebral Disk ◽

Key Factors ◽

Sixth Edition ◽

Whole Person ◽

Class 1 ◽

A Cell ◽

Intervertebral Disk Herniation ◽

The Individual

Abstract Ratings for “non-specific chronic, or chronic reoccurring, back pain” are based on the diagnosis-based impairment method whereby an impairment class, usually representing a range of impairment values within a cell of a grid, is selected by diagnosis and “specific criteria” (key factors). Within the impairment class, the default impairment value then can be modified using non-key factors or “grade modifiers” such as functional history, physical examination, and clinical studies using the net adjustment formula. The diagnosis of “nonspecific chronic, or chronic reoccurring, back pain” can be rated in class 0 and 1; the former has a default value of 0%, and the latter has a default value of 2% before any modifications. The key concept here is that the physician believes that the patient is experiencing pain, yet there are no related objective findings, most notably radiculopathy as distinguished from “nonverifiable radicular complaints.” If the individual is found not to have radiculopathy and the medical record shows that the patient has never had clinically verifiable radiculopathy, then the diagnosis of “intervertebral disk herniation and/or AOMSI [alteration of motion segment integrity] cannot be used.” If the patient is asymptomatic at maximum medical improvement, then impairment Class 0 should be chosen, not Class 1; a final whole person impairment rating of 1% indicates incorrect use of the methodology.

Download Full-text