Brain MRI for Knee OA

Chronic pain and knee osteoarthritis - context and gap

Chronic pain is a debilitating concern that can affect the health, employment, and lifestyle of individuals and puts a tremendous strain on the healthcare system. The prevalence of osteoarthritis (OA) is estimated at over 650 million worldwide, and people with knee OA account for 80% of the total burden (Cui et al. 2020). Globally in 2017, knee OA had a prevalence of 263,086,500 and incidence of 12,888,400 (James et al. 2018). A recent study also found that between 19%-43% of adults 40 years and older had knee OA markers, detected using magnetic resonance imaging (MRI), but were asymptomatic (Culvenor et al. 2019). It is estimated that more than 10 million Canadians will have knee OA by the year 2040 (Bombardier et al. 2011) significantly contributing to OA healthcare costs of 1.45 trillion (Bombardier et al. 2011; Sharif et al. 2017). The severity of pain and its impact on function have been identified by people living with knee OA as two of the most important effects of the disease (Smith et al. 2014), contributing to reduced quality of life (Neogi et al. 2013). Uncontrolled OA joint pain is associated with increased healthcare use, therefore improving pain management for OA in Canada may result in a significant healthcare expenditure savings of up to $488 billion (Bombardier et al. 2011). The pain experience and its associated mechanisms in people with knee OA are known to be complex, with alterations in pain signalling (Fingleton et al. 2015) a major factor in susceptibility to the development of persistent pain (Carlesso et al. 2019). Changes in nervous system sensitization do not necessarily correlate with changes in pain, demonstrating that other factors modulating pain, a top-down process, are at play (Skou et al. 2016; Edwards et la. 2016). Interestingly, between 10-20% of knee OA patients continue to experience intense pain even after a total knee replacement indicating that mechanisms beyond those involving joint structures are likely contributing to a complex pain process (Baker et la. 2007; Lundblad et al. 2008; Rice et al. 2018). In a disease where the joint is viewed as an organ in failure (Loeser et al. 2012), phenotyping is necessary to make sense of the heterogeneity. Specific to knee osteoarthritis (OA), pain is associated with changes in the joint structure due to inflammation (e.g., synovitis)(Dainese et al. 2022) or bone marrow lesions (Aso et al. 2021), but the underlying mechanisms are not fully understood (Gwilym et al. 2008).

Nociceptive, neuropathic, and nociplastic pain

In general, pain of known origin falls under two categories based on the cause: (i) nociceptive: pain caused by damage to a bodily tissue; (ii) neuropathic: pain caused by damage to a nerve or nervous system. However, chronic pain that is present without discernable nociceptive or neuropathic components is classified as nociplastic pain (IASP 2022). A recent study by Kosek et al. (2021) developed a simple scale for nociplastic pain diagnosis for clinical pain assessment based on i) the duration/regionality/type of pain, ii) history of hypersensitivities, iii) comorbidity identification, and iv) evoked pain hypersensitivity (Kosek et al 2021). As a visual representation, the Michigan Body Map (MBM) has been validated as a rapid and reliable tool for patients to identify general bodily regions where they experience chronic pain (Brummett et al. 2016). The MBM is a useful and simple tool to determine if pain is localized (i.e., likely nociceptive or neuropathic pain) or widespread (i.e., likely nociplastic pain)(Brummett et al. 2016). To address the unknown chronic pain mechanisms and pain presenting beyond nociceptive or neuropathic injuries, research has been focused on the theory of central sensitization as a phenomenon of nociplastic pain (Fitzcharles et al. 2021; Walsh 2021). The concept of central sensitization is that dysfunction in the central nervous system results in hypersensitivity to noxious and innocuous stimuli (de Boer et al. 2019; Neblett et al. 2013).

Current methods for nociplastic pain assessment

One method being used to examine nociceptive pain is functional magnetic resonance imaging (fMRI) data to measure the neurologic pain signature (NPS)(Han et al. 2022; Wager et al. 2013). The data collected from fMRIs is the brain's blood-oxygen level dependent (BOLD) signal which is closely connected to neurophysiological activation of the brain, and the signal fluctuates based on diamagnetic properties of deoxygenated blood and changes to blood flow and volume to activated or deactivated parts of the brain (Ogawa et al. 1990). The NPS has been shown to be a robust metric for quantitatively assessing the brain's activation in response to pain (Han et al. 2022; Wager et al. 2013). Based on the NPS, humans have several brain regions with key involvement in nociceptive pain interpretation that have proportionately increased or decreased activity relative to an individual's pain (Wager et al. 2013). The activation pattern of the NPS has been shown to be responsive to different forms of quantitative sensory testing (QST) such as thermal, mechanical (Wager et al. 2013), and electrical perturbations (Choi et al. 2011; Rütgen et al. 2015).

In women with chronic joint pain, there is evidence of cortical thinning related to chronic pain that exceeds that of normal aging, especially in the hippocampus, thalamus, and anterior cingulate cortex (de Kruijf et al. 2016). Within the context of OA, several studies have explored the effects of knee OA on fMRI activation changes (López-Solà et al. 2022; Pujol et al. 2017; Yue et al. 2018; Zunhammer et al. 2018). In summary, the studies indicate the BOLD signal is regionally elevated in patients with knee OA, but can be reduced by non-steroidal anti-inflammatory drugs (López-Solà et al. 2022) and analgesics (Yue et al. 2018). Spreading pain and heightened widespread pain sensitivity with corresponding brain abnormalities are commonly found in patients with knee OA (Pujol et al. 2017), and these pain-related brain abnormalities are distinct from potential psychological biases of placebos (Zunhammer et al. 2018).

To address the challenges of identifying the underlying pain mechanisms and providing patient-focused results, our research group has developed an analysis method for resting state fMRI (rsfMRI) and diffusion tensor imaging (DTI) that applies a personalized analysis. Based on the previous research using the NPS, the investigators expect to find regional brain alterations caused by pain that could improve our understanding of chronic pain that are also unique to each patient with knee OA. With our novel analysis, each patient with knee OA will have their MRI data compared to over 88 age- and sex-matched controls and a Z-scoring methodology can identify and objectively quantify abnormal brain regions. This can be applied as a compliment to clinical pain tests and more standard group-wise MRI analyses such as the NPS and functional networks.

While MRI has been useful in indicating brain regions altered by pain, inflammation and excitotoxicity may be underlying pathophysiologies behind nociplastic pain. A possible mechanistic pathway is that of kynurenine, having been linked to several inflammatory conditions, that can be neuroprotective if metabolized into kynurenic acid but neurotoxic if metabolized into quinolinic acid (Savitz 2020; Schwarcz and Stone 2017). Kynurenine is a metabolite of tryptophan and is detectable peripherally using the 13C-Tryptophan Breath Test (13C-TBT)(Teraishi et al. 2015). Although kynurenine has been researched for depression (Teraishi et al. 2015), there is mechanistic overlap at N-methyl-D-aspartate (NMDA) receptors between chronic pain and depression (Savitz 2020; Teraishi et al. 2015; Brandl et al. 2022). Two metabolomic studies of OA patients found decreased tryptophan concentrations (Abdelrazig et al. 2021; Van Pevenage et al. 2022), thus forming the possibility that the kynurenine pathways may be a useful bioindicator for OA.

Our solution for chronic nociplastic pain assessment

This proposed study will be a quantitative examination of nociplastic pain in patients with knee OA. Therefore, the primary objective for this study is to determine if chronic pain, more specifically chronic nociplastic pain, causes neuroplastic brain alterations. It is hypothesized that people with localized chronic pain (OA-Knee) and widespread chronic pain (OA-Knee+Body) exhibit abnormal brain function (e.g., network connectivity and ROI complexity) and reduced microstructural integrity of brain white matter (e.g., decreased fractional anisotropy), that would also be proportionately associated with pain sensitivity and mental state, in comparison to similarly aged individuals with knee osteoarthritis without chronic pain.

The cohort of participants will be separated into three patient groups: the control group will include participants with knee OA but no pain (Osteoarthritis - pain free (OA-PF)), the first exploratory group would include participants with knee OA with pain at the knee (OA-Knee), and the second exploratory group, indicative of chronic nociplastic pain would include people with knee OA with knee and widespread body pain (OA-Knee+Body). Participants in the OA-Knee+Body group will be representative of chronic pain patients presenting with nociplastic pain based on several clinical pain tests (e.g., Nociplastic pain criteria (Kosek et al. 2021), Michigan Body Map (Brummett et al. 2016), knee and wrist QST). Widespread bodily pain will be determined by identifying pain in >3 body regions on the Michigan Body Map and experiencing pain the knee and wrist pain pressure test. This study is designed to acquire resting state functional MRI (rsfMRI) and diffusion tensor imaging (DTI) MRI data at baseline and after experimentally induced pain (i.e., Cold Pressor Gel Test (Lapotka et al. 2017)), and compare the MRI data to clinical pain test results. The exploratory outcomes are to examine if there neuroinflammatory group differences based on the results of the 13C-TBT data.

Study Design

Study participants will be asked to complete a screening appointment over the phone prior to a data collection visit. Prospective participants will be notified of our study by their rheumatologist (i.e., Dr. Adachi) and will contact Dr. Danielli via email if they are interested in participating. Prospective participants will then contact Dr. Danielli to arrange a phone call to verbally obtain consent and perform the screening questions to determine eligibility. The screening visit will take about 30 minutes with Dr. Danielli. As part of the screening process prospective participants will be asked to complete the Nociplastic Pain Criteria (Kosek et al. 2021), the DSM-V Alcohol Use Disorder questionnaire, provide demographic information (i.e., to have balanced age and sex recruitment), and identify possible MRI contraindications. If someone is eligible to participate, an appointment will be then booked for the data collection study visit. All prospective participants will have been diagnosed with knee OA by their rheumatologist and the screening visit will be done to categorize participants into one of three study groups and efficiently book a data collection visit.

Informed written consent will be obtained in accordance with the World Medical Association Declaration of Helsinki prior to any data collection at the data collection study visit. All participants will receive the same data collection protocols. Depending on the preference of the participants, screening and data will be collected at either the Imaging Research Centre (IRC) at St. Joseph's Healthcare Hamilton or the Joint Department of Medical Imaging (JDMI) at Toronto Western Hospital. The questionnaires (i.e., the Depression Anxiety Stress Scale (DASS42)(Crawford et al. 2003), Intermittent and Constant Osteoarthritis Pain (ICOAP)(Hawker et al. 2008; Hawker et al. 2011), and Brief Pain Inventory (BPI)(Cleeland et al. 1991)), and participants will also have a wrist and knee pressure pain threshold tests; these will take about 15 minutes to complete. The baseline and secondary MRI acquisitions would take about 45 minutes each. Participants will be given 5-10 minutes between MRI scan sessions to take a break and stretch their legs. The data collection study visit will last about 2 hours.