Computed tomography Hounsfield unit values as a treatment response indicator for spinal metastatic lesions in patients with non-small-cell lung cancer: a retrospective study in Japan
Article information
Abstract
Study Design
A retrospective study.
Purpose
This study aimed to determine the impact of increased Hounsfield unit (HU) values for metastatic spinal lesions measured via computed tomography on the overall survival of patients with non-small-cell lung cancer (NSCLC) and identify factors associated with increased HU values in metastatic spinal lesions.
Overview of Literature
Previous studies have underscored the utility of the HU as a marker of treatment response in metastatic bone lesions. However, no prior studies have explored the relationship between HU changes in response to treatment and overall survival in patients with NSCLC.
Methods
This study included a total of 85 patients between 2016 and 2021. Nonsurgical treatments were provided by the respiratory medicine department. HU values for metastatic spinal lesions were evaluated upon diagnosis of spinal metastasis (baseline) and at 3, 6, and 12 months thereafter. Patients were then divided into two groups based on the median HU increase from baseline to 3 months. Overall survival was assessed using the Kaplan-Meier method.
Results
Based on the median change in HU value (124), 42 and 43 patients were categorized into the HU responder and non-responder groups, respectively. The median overall survival was significantly longer in the HU responder group than in the HU non-responder group (13.7 months vs. 6.4 months, p<0.001). Multiple linear regression analysis revealed that the use of antiresorptive agents and molecularly targeted therapies were factors significantly associated with an increase in HU.
Conclusions
An increase in HU values for metastatic spinal lesions after 3 months of treatment was correlated with a significantly longer overall survival in patients with NSCLC. Thus, HU measurements may not only serve as an easy and quantitative approach for evaluating treatment response in metastatic spinal lesions but also predict overall survival.
Introduction
Lung cancer remains one of the most prevalent malignancies worldwide, with approximately 21% of the cancer-related deaths in 2023 having been attributed to this condition [1]. In fact, studies have reported that 30%–40% of patients with advanced lung cancer develop bone metastases throughout the course of their disease [2,3] and that 65% of these metastases are identified during the initial staging process [4]. Several treatment options for metastatic lung cancer are available, including radiation therapy, bisphosphonates, and anticancer drugs [5]. In particular, the development of new drugs, including molecular targeted therapies and immunotherapy, has dramatically changed the systemic therapy for metastatic non-small-cell lung cancer (NSCLC) [6,7].
The Hounsfield unit (HU), a measurement used for evaluating the bone mineral density based on computed tomography (CT), has been widely applied in the field of spinal medicine, with support from recent review studies [8,9]. Conventional metrics used in assessing the therapeutic response in patients with spinal metastasis involve symptom palliation and pain [10,11]. However, relying on palliative measures introduces subjectivity to the evaluation of treatment responses. Consequently, one review article has recommended the use of HU values for the quantitative evaluation of treatment responses in metastatic bone disease [12]. Although prior studies have highlighted the utility of the HU as a marker of treatment response in bone metastatic lesions, its value has been limited due to the confounding impact of systemic therapies and various types of cancers [10,11,13]. In line with this, no previous studies have investigated the relationship between changes in HU values in response to treatment and overall survival in patients with NSCLC.
The current study therefore aimed to determine the impact of increased HU on the overall survival of patients receiving systemic therapy for NSCLC and identify factors affecting the increase in HU values for metastatic spinal lesions.
Materials and Methods
Study design and ethics statement
The protocol for this retrospective cohort study was approved by the institutional review board (IRB) of Yodogawa Christian Hospital (IRB no., 2023-072). All procedures were performed in accordance with the ethical principles stated in the Declaration of Helsinki. The requirement for informed consent from individual patients was omitted because of the retrospective design of this study.
Patients
This study included a total of 146 patients diagnosed with spinal metastases from NSCLC at Yodogawa Christian Hospital between 2016 and 2021. Spinal metastases were diagnosed radiologically using CT, magnetic resonance imaging, or positron emission tomography. Patients who failed to undergo follow-up CT after having been diagnosed with bone metastasis due to death within 2 months (n=33) or those who had incomplete radiological data on follow-up CT at 3 months (n=28) were excluded from analysis. The minimum follow-up duration after diagnosis with bone metastasis was 6 months, except for those who died within the first 6 months. Finally, 85 patients (including 42 women) were included in this study. The average age of the patients was 70.2±9.1 years upon diagnosis with spinal metastasis. The respiratory medicine department provided nonsurgical treatments, including chemotherapy, antiresorptive agents, and general care. When necessary, the respiratory medicine department consulted with the radiology and orthopedic departments regarding radiation therapy and spinal surgery, respectively.
Patient demographic data
Patient demographic data, including age, sex, histology reports, bone lesion type, modified Katagiri score, primary site treatment (molecularly targeted therapy or immune checkpoint inhibitors), use of antiresorptive agents, radiation therapy, and spinal surgery, were extracted from the medical records. The modified Katagiri score is a scoring system that incorporates the use of molecularly targeted drugs to determine the prognosis of patients with metastatic bone tumors [14]. Patients were scored from 0 to 10 and were subsequently divided into the following three groups: low-risk (0–3), intermediate-risk (4–6), and high-risk (7–10) groups.
Measurement of HU values
As part of their routine follow-up, all patients underwent multislice CT on an Aquilion PRIME CT scanner (Canon Medical Systems Co., Tochigi, Japan). HU for the metastatic and nonmetastatic lesions were measured separately to ensure reliability and reproducibility. For metastatic spinal lesions, the most representative slice among the baseline images from each patient was selected to measure the HU. If the most severe condition could not be determined at baseline, the level showing the most sclerotic changes from baseline to 3 months was selected. Bone lesions (lytic, mixed, and sclerotic) were assessed using the method shown in Fig. 1. For nonmetastatic spinal lesions, the HU was determined by placing an elliptical region of interest in the axial images at L1–L4 in the middle of the vertebral body. The HU was determined from the average value measured at each level, in accordance with Schreiber et al. [15] who recommended drawing the largest possible elliptical region of interest, excluding the cortical margin. HU values were evaluated for both metastatic and nonmetastatic lesions upon diagnosis of spinal metastasis (baseline) and at 3, 6, and 12 months thereafter. Owing to the retrospective nature of this study, a 1-month difference at each time point was allowed at 3, 6, and 12 months after the diagnosis of spinal metastasis.
Region of interest setting of computed tomography images of the metastatic lesions at diagnosis (baseline) of bone metastasis and after treatment. (A) Lytic bone metastasis. (B) Mixed-type bone metastasis. (C) Sclerotic bone metastasis.
Statistical analysis
Patients were classified into two groups, namely the HU responder and non-responder groups, based on changes in HU values for metastatic spinal lesions 3 months after the diagnosis of spinal metastasis. The median change in HU value from baseline 3 months after diagnosis of spinal metastasis was set as the cutoff value for the increase in HU. Demographic data for the two groups were compared using univariate analysis, with Fisher’s exact test and the Mann-Whitney U test being used for categorical and continuous variables, respectively. HU values for metastatic and nonmetastatic spinal lesions recorded at baseline and 3, 6, and 12 months after diagnosis of spinal metastasis were compared using a mixed-effects model. Overall survival was estimated using the Kaplan-Meier method. The log-rank test and Cox proportional hazards regression model adjusting for age, sex, and modified Katagiri score were used for univariate and multivariate analyses, respectively. Patients who were lost to follow-up were censored at the date of last contact/follow-up. Overall survival was calculated from the date of bone metastasis diagnosis to the date of death. Finally, multiple linear regression analyses were performed to examine the association between HU values for metastatic spinal lesions recorded 3 months after diagnosis and the type of treatment and bone lesions. Age, sex, baseline HU value for the metastatic lesion, use of antiresorptive agents, radiation therapy, molecular targeted therapy, immune checkpoint inhibitors, and bone lesion type (sclerotic, lytic, and mixed) were included as explanatory variables. All analyses were performed using R software ver. 3.5.1 (The R Foundation for Statistical Computing, Vienna, Austria), with p<0.05 indicating statistical significance.
Results
Clinical course of HU
The changes in HU values for metastatic lesions and nonmetastatic bones are summarized in Table 1. Overall, the HU values for the metastatic lesions increased significantly throughout the study period (p<0.001). The median increase in HU value between baseline and 3 months after diagnosis was 124.0. Based on a cutoff value of 124, 42 and 43 patients were categorized as HU responders and non-responders, respectively. A significant difference in the HU value for metastatic bone was observed between HU responders and non-responders 3 months (441.7 versus 237.2, p<0.001), 6 months (531.1 versus 325.5, p<0.001), and 12 months (656.0 versus 527.9, p<0.001) after diagnosis. No significant differences were observed between the two groups at baseline (189.9 versus 181.6, p=0.752). Although the HU value for nonmetastatic bone increased throughout the study period, no significant difference was observed overall (p=0.051) or between the two groups (p=0.226).
Changes in HU values of metastatic and non-metastatic bone lesions at various time points
Demographics
Table 2 summarizes the characteristics of the patient included in the two groups. The use of molecularly targeted therapy was significantly higher in the HU responder group than in the HU non-responder group (47.6% versus 20.9%, p=0.012). No significant differences in the use of antiresorptive agents (85.7% versus 74.4%, p=0.279) or in the proportion of denosumab (88.9% versus 81.3%) and bisphosphonate use (11.1% versus 18.7%) were observed between the two groups. Five patients underwent surgery for spinal metastasis due to neurological deficits, with no significant difference having been observed between the two groups (4.8% versus 7.1%, p=1.000).
Patient demographic data
Overall survival
Within 2 years after the diagnosis of spinal metastasis, a total of 31 deaths (74%) and 38 deaths (88%) occurred in the HU responder and non-responder groups, respectively (Fig. 2). The median overall survival time was significantly longer in the HU responder group than in the HU non-responder group (13.7 months versus 6.4 months, p<0.001). A stratified Cox regression model that adjusted for age, sex, and modified Katagiri score showed that patients in the HU responder group had a significantly decreased likelihood of death (hazard ratio, 0.52; 95% confidence interval [CI], 0.32–0.83; p=0.006).
Overall survival of the Hounsfield unit (HU) responder group versus the non-responder group. CI, confidence interval; HR, hazard ratio.
Factors associated with an increased HU value for metastatic spinal lesions
Multiple linear regression analysis, adjusting for age, sex, and baseline HU value, revealed that the use of antiresorptive agents (β=89.7, p=0.010) and molecularly targeted therapy (β=92.3, p=0.017) were the factors significantly associated with an increase in the HU value for metastatic spinal lesions 3 months after the diagnosis of bone metastasis (Table 3).
Multiple linear regression analysis of factors associated with the HU value of metastatic bone lesions after treatment for 3 months
Discussion
The current study established a significant correlation between a substantial increase in HU value for metastatic spinal lesions after 3 months of treatment for NSCLC and markedly prolonged overall survival, with a stratified Cox regression model showing a concomitant reduction in the mortality rate. Furthermore, the use of antiresorptive agents and molecularly targeted therapies were identified as factors significantly associated with increased HU values for metastatic spinal lesions after a 3-month intervention period. The current study has been the first to report that changes in HU in response to the treatment of metastatic lesions in patients with NSCLC influences overall survival. Moreover, this study has been the first to show that molecular targeted therapies were significantly associated with an increase in HU. The results of our analysis support the notion that HU measurements may serve as an objective and quantitative approach for evaluating treatment response in patients with metastatic spinal lesions.
One previous study did attempt to measure HU values as a marker of treatment response in bone metastatic lesions [13]. Furthermore, a review article concluded that this method is a reproducible, reliable, and quantifiable marker for measuring therapeutic responses in patients with metastatic bone disease [12]. In the present study, the HU responder and non-responder groups were stratified based on changes in HU values for metastatic lesions from baseline after 3 months, with our findings showing that significant differences persisted until 6 and 12 months (p<0.001). Nonmetastatic lumbar HU values, which served as controls, did not significantly differ throughout the study period. Despite the potential bone loss associated with some cancer treatments [16,17], our findings suggest that treatment, including systemic therapy and antiresorptive agents for NSCLC, positively influenced the recalcification of metastatic spinal lesions for at least 1 year without inducing significant bone loss at nonmetastatic sites.
The current study revealed that the overall median survival time was significantly longer in the HU responder group than in the HU non-responder group (13.7 months versus 6.4 months, p<0.001). Stratified Cox modeling adjusting for age, sex, and modified Katagiri score demonstrated a substantial decrease in the death rate among HU responders (hazard ratio, 0.52; p=0.006). Scoring systems, such as Katagiri or Tokuhashi, have been used to evaluate the prognosis and survival of patients with metastatic spinal cancer [12,14], with a recent review underscoring their reliability and usefulness [18]. Although these scoring systems are essential for evaluating the overall survival upon treatment initiation, observing treatment responses throughout the clinical course is also crucial. Nakata et al. [19] reported that the early response to treatment for bone metastases among patients with NSCLC, as measured by the MD Anderson response criteria, predicted a better prognosis in such patients. Therefore, our results suggest that the measurement of HU values in patients with metastatic spinal lesions can be an easy, quick, and readily available approach for the early prediction of treatment response and its impact on overall survival.
Our results also revealed that the use of antiresorptive agents and molecular targeted therapies significantly improved HU values for metastatic spinal lesions after 3 months of treatment. Indeed, a randomized controlled trial had previously demonstrated the clinical efficacy of antiresorptive agents for skeletal metastases in patients with NSCLC [20]. Additionally, a recent review by Cadieux et al. [21] supported the safety and efficacy of antiresorptive therapy in patients with metastatic cancer after reviewing available data from the 10 years after their approval. Therefore, the observed correlation between the use of antiresorptive agents and improvement in HU appears reasonable. Although the impact of molecular targeted therapy on metastatic spinal lesions remains unknown, previous reviews and studies have suggested its clinical efficacy [6,7]. Interestingly, Hu et al. [22] indicated that radiotherapy combined with targeted therapy strongly inhibited cyclooxygenase 2 and vascular endothelial growth factor expression in bone metastasis from lung cancer, which subsequently improved treatment efficacy and prolonged survival. Thus, molecular targeted therapy may exert a strong influence on the treatment of metastatic spinal lesions, potentially reducing bone resorption by inhibiting osteoclast activity and contributing to the increase in HU values for these lesions. Notably, immune checkpoint inhibitors did not significantly improve the HU values for metastatic spinal lesions. Although a few studies have suggested the effectiveness of immune checkpoint inhibitors in metastatic bone disease, Nakata et al. [19] claimed that bone response to nivolumab monotherapy was significantly correlated with tumor control and was useful apparat for predicting prognosis among patients with NSCLC who had bone metastases. Therefore, further studies are required to assess the effects of immune checkpoint inhibitors on metastatic spinal lesions in patients with NSCLC.
This study has several strengths. First, this study only included patients with NSCLC who had spinal metastasis. Moreover, the inclusion period was after immune checkpoint inhibitors had been approved in 2015 given the dramatic change in the treatment for NSCLC. Additionally, the presence of a hospice care facility in the hospital ensured treatment continuity until the day of death, minimized loss to follow-up, and enhanced the reliability of overall survival results in the HU responder group.
This study has some limitations worth noting. First, our definition of HU responders relied on observed improvements in HU values from baseline after 3 months of treatment, which may limit its applicability as a scoring system for assessing the overall prognosis at the time of spinal metastasis diagnosis. Second, the measurement of HU values could have been subjective. To address this concern, we ensured uniformity in the measurement of metastatic lesions by following a consistent approach based on the types of bone lesions as shown in Fig. 1, thereby minimizing subjectivity. Further studies are needed to reevaluate these results using a larger sample.
Conclusions
In conclusion, the current study revealed a correlation between an increase in HU values for metastatic spinal lesions after 3 months of treatment for NSCLC and significantly prolonged overall survival. The use of antiresorptive agents and molecular targeted therapies were also identified as factors significantly associated with HU improvement. As such, HU measurement may not only serve as an easy and quantitative approach for evaluating treatment response in metastatic spinal lesions but may also predict overall survival in patients with NSCLC.
Key Points
No prior studies have investigated the relationship between Hounsfield unit (HU) changes in response to treatment and overall survival in patients with non-small-cell lung cancer (NSCLC)
This study found a correlation between increased HU values for metastatic spinal lesions after 3 months of treatment for NSCLC and significantly prolonged overall survival
HU measurements may serve not only as a quantitative method for evaluating treatment response in metastatic spinal lesions but also as a potential predictor of overall survival in patients with NSCLC.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Author Contributions
HT and SD have given substantial contributions to the study conception and design, data analysis and drafting the manuscript. RS, KT, MH, and HN contributed to the critical revision of the manuscript for important intellectual content. All authors read and approved the final version of the manuscript.