BACKGROUND:Healthcare professionals are increasingly providing medical information to patients. However, few objective evaluations of the information's usefulness exist. This study aimed to evaluate patient satisfaction with medical information sources, including the JLCS Guidebook and patient advocacy groups, using the European Organization for Research and Treatment of Cancer's (EORTC) Quality of Life 25-item information module (QLQ-INFO25). METHODS:Patients with thoracic malignancies were enrolled in the trial in response to calls from patient support groups and the Japanese Lung Cancer Society. Patient satisfaction was based on information obtained from the 2019 Japanese Lung Cancer Society's patient guidebook and a web-based survey of patients with lung cancer. Satisfaction with the information was analyzed using EORTC QLQ-INFO25 scores and stratified by guidebook use (users vs. non-users). RESULTS:Of 332 participants (110 male, 222 female), 283 (85.2%) completed the questionnaire, among which 82 (29.0%) read the guidebook. Participants found the research through online sources or by responding to an invitation from their primary physician or patient support group. Seventy-one users (91%) were satisfied with the guidebook content, as indicated by their individual QLQ-INFO25 scale scores. No statistically significant differences were observed between guidebook users and non-users; however, a trend toward higher satisfaction was noted among participants who used both the guidebook and advocacy groups. CONCLUSIONS:Our findings suggest that guidebooks and advocacy groups may contribute to improved patient satisfaction; however, further controlled studies are warranted. Our results provide an objective basis for discussing information provision for patients with lung cancer. Strategies should be developed to communicate information reflecting patients' diverse backgrounds.

J-TAIL-2 evaluated the efficacy and safety of atezolizumab, an anti-programmed death-ligand 1 (PD-L1), plus chemotherapy in patients with non-small cell lung cancer (NSCLC) and extensive-stage small cell lung cancer (ES-SCLC) in Japanese clinical practice, demonstrating comparable outcomes to those in the corresponding Phase 3 trials. Although PD-L1 expression is predictive of atezolizumab efficacy in some settings, additional minimally invasive, blood-based biomarkers are needed. This exploratory biomarker study included 359 J-TAIL-2 patients. The NSCLC cohort received atezolizumab combined with carboplatin and nab-paclitaxel (CnP, n = 42), cisplatin/carboplatin and pemetrexed (n = 72), or bevacizumab plus carboplatin and paclitaxel (bev + CP, n = 135). The ES-SCLC cohort received atezolizumab plus carboplatin and etoposide (n = 100). Approximately 560 cancer- or immune-related proteins were evaluated using the Proximity Extension Assay from blood plasma collected at three timepoints: baseline, before the second atezolizumab dose, and at the onset of immune-related adverse events (irAEs). Protein-wise comparisons were made to evaluate significant changes (P < 0.05 and > 0.5 log2 fold change) and linked to clinical outcomes reported in J-TAIL-2. IL-6, MUC-16, and KRT-19 were associated with shorter progression-free survival across multiple regimens. Granzyme A and B, immune activation markers, were elevated in responders who received atezolizumab + CnP and atezolizumab + bev + CP. High levels of baseline immune stimulation proteins were associated with irAEs across regimens. Protein expression changes were varied and regimen-dependent in subgroup analyses including older patients and those with EGFR-mutant tumors. These data warrant further investigation across different cancer types and atezolizumab-containing regimens to determine their relevance to efficacy and irAE occurrence of atezolizumab combination therapy. Trial Registration: ClinicalTrials.gov ID, NCT04501497 (J-TAIL-2) and NCT04818983 (biomarker study).

ABSTRACT Introduction Interstitial lung disease (ILD) is a concerning adverse event associated with immune checkpoint inhibitors, including durvalumab. This analysis of the multicenter AYAME study assessed the longterm safety of durvalumab, focusing on ILD. Methods AYAME enrolled patients prescribed durvalumab for unresectable stage III non‐small cell lung cancer (NSCLC) after chemoradiotherapy (CRT) from July 2019 to December 2020 in Japan. Patients received durvalumab for ≤ 12 months and were prospectively followed for 3 years. Incidence, severity, and management of ILD were examined. Multivariable logistic regression analysis was conducted to investigate the association of patient characteristics with grade ≥ 2 ILD occurrence. Results Of 511 patients in the safety analysis population, ILD occurred in 383 patients (75.0%) from durvalumab initiation to subsequent treatment start; median time to occurrence was 44.0 days. ILD led to permanent durvalumab discontinuation in 121/383 patients (31.6%), dose interruption in 111/383 (29.0%), and corticosteroid intervention in 168/383 (43.9%). Grade ≥ 2 ILD occurrence was higher in patients with volume of lung parenchyma that received 20 Gy (V20) ≥ 25% versus < 25% (60.3% vs. 29.4%; adjusted odds ratio 2.20, 95% confidence interval 1.08–4.48), 5 Gy (V5) ≥ median (37.85%) versus < median (51.5% vs. 27.2%; 1.79, 1.07–3.00), and grade 1 ILD before durvalumab administration versus no ILD (50.0% vs. 37.6%; 2.11, 1.07–4.17). Conclusions This study of durvalumab for NSCLC after CRT in a real‐world setting indicated factors associated with grade ≥ 2 ILD and provided valuable insights into ILD management strategies. These findings may contribute to establishing effective approaches to minimize ILD risk in practice. Trial Registration: UMIN000037090/NCT03995875

PURPOSE:Durvalumab therapy following concurrent chemoradiotherapy (cCRT) improves progression-free survival (PFS) in patients with unresectable locally advanced non-small cell lung cancer. In this prospective observational study, we evaluated the changes in peripheral blood immune cell counts to elucidate the immunologic mechanisms underlying cCRT and durvalumab therapy. EXPERIMENTAL DESIGN:Peripheral blood mononuclear cell (PBMC) samples were collected at four time points: before cCRT, after cCRT, at the start of durvalumab, and 8 weeks after the start of durvalumab, and analyzed by multicolor flow cytometry. RESULTS:Of the 149 enrolled patients, 115 received durvalumab consolidation therapy after cCRT. The median PFS in the overall population was 24.2 months, and the 3-year PFS rate was 38.9%. PBMC analysis showed an increased effector fraction of CD4+ T cells before and after cCRT but no change in CD8+ T cells. Following durvalumab therapy, the effector fraction ratio of CD8+ T cells (CD62Llow CD8+ T cells) increased and positively correlated with increased CD62Llow CD4+ T cells during cCRT. Patients whose proportion of CD62Llow CD4+ T cells exceeded the threshold for cCRT had better PFS than those below the threshold. Patients whose CD62Llow CD8+ T-cell proportion exceeded the threshold after durvalumab therapy showed prolonged PFS compared with those below the threshold. CONCLUSIONS:cCRT promotes an increase in effector CD4+ T cells, and the subsequent increase in CD8+ T cells following durvalumab therapy prolongs PFS. Peripheral blood effector-type CD4+ and CD8+ T cells are potential biomarkers for evaluating the immune status of patients and predicting treatment efficacy.

We investigated immunologic biomarkers that predict outcomes of patients with previously treated metastatic non-small cell lung cancer who were enrolled in the J-TAIL study, a prospective, observational study of atezolizumab monotherapy. Of 262 patients participating in the J-TAIL exploratory study, peripheral blood mononuclear cells were obtained from 51 patients and analyzed by T-cell fractionation analysis using Helios mass cytometry and serum proteomics analysis. Following treatment with atezolizumab, an increase in programmed cell death-1 (PD-1)-expressing CD8 and CD4 T-cell populations was observed. A more pronounced increase in PD-1 expression was seen in T cells from patients whose progression-free survival (PFS) was 100 days or longer compared with those with shorter PFS. The proximity extension assay, which is highly sensitive multiplex analysis technology that combines antibody-based affinity assays with next-generation sequencing, showed a significant increase in FOXO1, possibly in response to precursor-exhausted T-cell population activation. Immune-related adverse events were associated with a high percentage of PD-1-positive cells on effector memory CD8 T cells, which was thought to be accompanied by extremely high CD8 T-cell activation. Further analysis distinguished poor prognosis populations with significant differences in CD62Lhigh Th7R and CXCR3+ component of Th7R (CXCR3+ Th7R) within the population with PFS < 50 days. Patients with low Th7R or CXCR3+ Th7R percentages prior to atezolizumab treatment had significantly poorer overall survival. These findings provide valuable insights regarding T-cell kinetics and biomarkers in atezolizumab therapy and may offer promising directions for future research. Trial Registration: UMIN Clinical Trials Registry: UMIN000033133 and UMIN000035567; ClinicalTrials.gov: NCT03645330.

Pivotal phase 3 trials leading to the approvals of atezolizumab─chemotherapy combinations for non-small cell and extensive-stage small cell lung cancer (NSCLC and ES-SCLC) had strict eligibility criteria. J-TAIL-2 was a prospective, observational study in Japan that evaluated atezolizumab regimens for advanced NSCLC/ES-SCLC, including patients ineligible for global trials. The primary endpoint was 12-month overall survival (OS); safety and efficacy in subgroups defined by age, ECOG PS and/or G8 score, and creatinine clearance were key secondary endpoints. As of February 3, 2023, 1217 patients were treated in clinical practice based on Japanese labeling/treatment guidelines. Patients received atezolizumab with either carboplatin and nab-paclitaxel (atezo + CnP), carboplatin/cisplatin and pemetrexed (atezo + PP), or bevacizumab and carboplatin and paclitaxel (atezo + bev + CP) in the NSCLC cohort (n = 814) or with carboplatin and etoposide (atezo + CE) in the ES-SCLC cohort (n = 403). Overall, 53.5% were ≥ 70 years old, and 11.8% had ECOG PS ≥ 2; median G8 scores were 13 (NSCLC cohort) and 12 (ES-SCLC). Patients < 70 and ≥ 70 years had similar median (m)OS and progression-free survival (mPFS) across treatment regimens. Patients with ECOG PS < 2 and G8 score ≥ median had the highest mOS/PFS vs. patients with ECOG PS ≥ 2 and G8 score<median. Patients ≥ 70 years had higher incidence of Grade ≥ 3 adverse events with atezo + CnP and atezo + PP than patients < 70 years; incidences were similar between groups for other regimens. Older vs. younger patients also had higher incidence of interstitial lung disease. Overall, these results suggest that atezolizumab-containing regimens remain effective in older patients, with no new safety signals. Trial Registration: ClinicalTrials.gov: NCT04501497.

Supplementary Figure S1. PFS according to the grade of irAE (A) 4 weeks landmark PFS of irAE according to the grade (B) 8 weeks landmark PFS of irAE according to the grade One event with undetermined grade was excluded from the analysis. Abbreviations: CI, confidence interval; HR, hazard ratio; irAE, immune-related adverse event; PFS, progression-free survival

Supplementary Table S6. Association between skin or endocrine disorder irAEs and predictors of ICI effect a Mann–Whitney U Test. b Patients who had progression within 4 or 6 weeks after the initiation of treatment were excluded. Abbreviations: CRP, C-reactive protein; ICI, immune checkpoint inhibitor; IHC, immunohistochemical; irAE, immune-related adverse event; NLR, neutrophil-to-lymphocyte ratio; PD-L1, programmed death ligand-1; Q, quartile; SD, standard deviation.

BACKGROUND:Japan has a higher incidence of drug-induced interstitial pneumonia (IP) than other countries. Understanding the clinical characteristics of drug-induced IP in Japan will facilitate accurate diagnosis of the disease. The aim of this study was to analyze and report the clinical characteristics of drug-induced IP in Japan. METHODS:We analyzed the characteristics of drug-induced IP diagnosed between 2009 and 2022 in Japan. Case data were collated with the cooperation of pharmaceutical companies, as requested by the Ministry of Health, Labour and Welfare. RESULTS:We identified 285 patients with drug-induced IP diagnosed in medical institutions. The most common causative drugs identified were antineoplastic agents (56.1 %). Common subjective symptoms included dyspnea (158 patients [55.4 %]), cough (132 patients [46.3 %]), and fever (115 patients [40.4 %]). In addition, 177 patients (73.1 % of those measured) showed elevated serum KL-6, SP-D, or SP-A levels. Approximately 60 % of the patients developed IP within 90 days from the initiation of drug therapy, whereas 15 % developed the disease after >1 year. Regarding imaging findings, 249 patients (87.4 %) showed bilateral shadows. Corticosteroid therapy was administered to 222 patients (77.9 %). The treatment outcomes of drug-induced IP in Japan were favorable, with improvement observed in 273 patients (95.8 %). CONCLUSIONS:Dyspnea, cough, and fever after administration of a drug; elevated serum KL-6, SP-D, and SP-A levels; and bilateral shadows were identified as clinical features of drug-induced IP in Japan. Drug-induced IP should be considered a differential diagnosis even if the patient started using the causative drug long ago.

Supplementary Table S2. Univariable and multivariable analysis of OS Abbreviations: ECOG PS, Eastern Cooperative Oncology Group performance status; HR, hazard ratio; ICI, immune checkpoint inhibitor; IHC, immunohistochemistry; irAE, immune-related adverse event; OS, overall survival; PD-L1, programmed death ligand-1; TPS, tumor proportion score.

INTRODUCTION:Based on the PACIFIC trial results, durvalumab after chemoradiotherapy (CRT) is the standard of care for unresectable stage III NSCLC. However, few large-scale studies have prospectively evaluated its long-term effectiveness and safety, including subsequent treatment. METHODS:This multicenter, noninterventional study (AYAME) enrolled patients who were prescribed durvalumab for unresectable stage III NSCLC after CRT (July 2019-December 2020; 52 sites; Japan). Patients received durvalumab for a maximum of 12 months and were prospectively followed up for 3 years, including the post-durvalumab treatment period. Primary end points were real-world progression-free survival (rwPFS), the incidence of interstitial lung disease (ILD), and adverse events of special interest (AESIs). RESULTS:Of the 529 enrolled patients, 512 received durvalumab and 511 comprised the safety analysis population. The median time to onset of the first presentation or occurrence of ILD was 45.0 days. ILD of any grade occurred in 387 patients (75.7%) over the 3-year study period: grade 3, 57 (11.2%); grade 4, none (0.0%); and grade 5, 9 (1.8%). The most common AESIs of any grade during durvalumab treatment were thyroid dysfunction (12.5%), hepatic dysfunction (7.2%), and colitis (2.0%). No durvalumab treatment-related ILD or AESIs occurred during the post-treatment period. The median [95% confidence interval] rwPFS was 23.2 [18.2, 27.2] months in the effectiveness analysis population (n = 495). rwPFS rates were 62.5%, 54.6%, 49.4%, and 40.4% at 12, 18, 24, and 36 months, respectively. CONCLUSIONS:This first and largest prospective, observational study demonstrated the long-term safety and effectiveness of durvalumab for unresectable stage III NSCLC after CRT in a real-world clinical setting. CLINICAL TRIAL REGISTRATION:UMIN000037090 and NCT03995875.

Supplementary Table S3. Univariable and multivariable analysis of PFS Abbreviations: ECOG PS, Eastern Cooperative Oncology Group performance status; HR, hazard ratio; ICI, immune checkpoint inhibitor; IHC, immunohistochemistry; irAE, immune-related adverse event; PD-L1, programmed death ligand-1; PFS, progression-free survival; TPS, tumor proportion score.

BACKGROUND:We conducted a randomize phase II study to evaluate the efficacy and safety of topoisomerase II inhibitor amrubicin plus topoisomerase I inhibitor irinotecan (AI) compared with cisplatin plus irinotecan (PI) as first-line therapy in patients with extensive-disease (ED) small-cell lung cancer (SCLC). PATIENTS AND METHODS:Chemo-naïve patients with pathologically proven ED-SCLC (including limited disease (LD) SCLC with malignant effusion) were enrolled. Patients were randomized 1:1 to receive either AI (amrubicin 90mg/m2 on day 1 and irinotecan 50mg/m2 on days 1 and 8 of a 21-day cycle) or PI (cisplatin 60mg/m2 on day 1 and irinotecan 60mg/m2 on days 1, 8 and 15 of a 28-day cycle). The primary endpoint was overall survival proportion at 1 year. RESULTS:A total of 100 patients were randomly assigned to AI (n = 50) or to PI (n = 50). The 1-year overall survival proportions were 68.0% (95% confidence interval (CI): 56.2-82.2%) for AI and 59.2% (46.9-74.7%) for PI (1-sided P = .18). Median survival time was 14.8 months for AI and 13.5 months for PI with a hazard ratio (HR) of 0.618 (0.398-0.961, stratified log-rank test P = .031). Median progression-free survival time was 4.8 months for AI and 5.4 months for PI (stratified log-rank test, P = .54). Objective response rate was 70.0% (55.4-82.1%) for AI and 55.1% (40.2-69.3%) for PI (Fisher exact test, P = .15). There was no significant difference in hematological toxicity, whereas rates of vomiting, loss of appetite, diarrhea, and elevated serum creatinine are more frequent in PI. Interstitial lung disease (Grade 2 or 3) developed in 5 patients in AI and in 1 patient in PI. There was no treatment-related death. CONCLUSION:Although the study did not meet its primary endpoint, AI showed promising efficacy and good tolerability in chemo-naïve patients with ED-SCLC.

Introduction:Effective predictors of response to atezolizumab plus carboplatin/etoposide (CE) therapy in extensive-stage SCLC (ES-SCLC) remain limited. This exploratory analysis from J-TAIL-2 aimed to identify markers of survival benefit with atezolizumab plus CE therapy in ES-SCLC. Methods:J-TAIL-2 (ClinicalTrials.gov ID, NCT04501497) was a multicenter observational study that enrolled patients receiving atezolizumab plus CE (ES-SCLC cohort) in clinical practice in Japan per local label and treatment guidelines. In this exploratory analysis, the association of CD8+ tumor-infiltrating lymphocyte (TIL) density and SCLC subtypes (SCLC-A [ASCL1 dominant], SCLC-N [NEUROD1 dominant], SCLC-P [ASCL1/NEUROD1 double-negative with POU2F3 expression], and SCLC-O [ASCL1/NEUROD1 double-negative not otherwise specified]) with overall survival (OS) and progression-free survival (PFS) was evaluated. SCLC subtyping was performed by immunohistochemistry. Results:SCLC samples (n = 100; data cutoff, February 3, 2023) were categorized as SCLC-A (73%), SCLC-N (16%), SCLC-P (8%), and SCLC-O (3%). Among 96 patients who received first-line atezolizumab plus CE, median age was 72 (range, 39-87) years and 81% were male. Furthermore, 56 patients were classified into the CD8+ TIL-high subgroup and 40 into the TIL-low subgroup. Median (m)PFS with atezolizumab plus CE was 6.1 months (95% confidence interval [CI]: 4.5-7.5) in the TIL-high versus 4.4 months (95% CI: 4.0-5.1) in the TIL-low subgroup (p = 0.01); mOS was 18.4 (95% CI: 11.8-not estimable) versus 10.8 months (95% CI: 7.7-16.2; p = 0.04). mOS and mPFS were not significantly different between SCLC subtypes but were numerically shorter in the SCLC-N group. Conclusions:CD8+ TIL density is a potential biomarker of clinical benefit in ES-SCLC and may facilitate patient selection for atezolizumab combination therapy.

Introduction Effective predictors of response to atezolizumab+carboplatin/etoposide (CE) therapy in extensive-stage small cell lung cancer (ES-SCLC) remain limited. This exploratory analysis from J-TAIL-2 aimed to identify markers of survival benefit with atezolizumab+CE therapy in ES-SCLC. Methods J-TAIL-2 (ClinicalTrials.gov ID, NCT04501497) was a multicenter observational study that enrolled patients receiving atezolizumab + CE (ES-SCLC cohort) in clinical practice in Japan per local label and treatment guidelines. In this exploratory analysis, the association of CD8+ tumor-infiltrating lymphocyte (TIL) density and SCLC subtypes (SCLC-A [ASCL1-dominant], SCLC-N [NEUROD1-dominant], SCLC-P [ASCL1/NEUROD1 double-negative with POU2F3 expression], and SCLC-O [ASCL1/NEUROD1 double-negative not otherwise specified] with overall survival (OS) and progression-free survival (PFS) were evaluated. SCLC subtyping was performed by immunohistochemistry. Results SCLC samples (n=100; data cutoff, February 3, 2023) were categorized as SCLC-A (73%), SCLC-N (16%), SCLC-P (8%), and SCLC-O (3%). Among 96 patients who received first-line atezolizumab + CE, median age was 72 years (range, 39–87) and 81% were male. Fifty-six patients were classified into the CD8+ TIL-high subgroup and 40 into the TIL-low subgroup. Median (m)PFS with atezolizumab + CE was 6.1 months (95% CI: 4.5-7.5) in the TIL-high vs 4.4 months (95% CI: 4.0-5.1) in the TIL-low subgroup (P=0.01); mOS was 18.4 (95% CI: 11.8-NE) vs 10.8 months (95% CI: 7.7-16.2; P=0.04). mOS and mPFS were not significantly different between SCLC subtypes but were numerically shorter in the SCLC-N group. Conclusions CD8+ TIL density is a potential biomarker of clinical benefit in ES-SCLC and may facilitate patient selection for atezolizumab combination therapy.

First-line atezolizumab combination therapies were approved for the treatment of metastatic non-small cell lung cancer (NSCLC) based on results from the global phase 3 trials IMpower130, IMpower132, and IMpower150. These trials reported 12-month overall survival (OS) rates of 60%-67% with atezolizumab combination therapy. J-TAIL-2 (NCT04501497), a prospective, multicenter, observational study, evaluated atezolizumab combination therapy in routine clinical practice in Japan. Patients ≥ 20 years old with NSCLC received atezolizumab plus carboplatin and nab-paclitaxel (atezo + CnP), atezolizumab plus carboplatin or cisplatin plus pemetrexed (atezo + PP), or atezolizumab plus bevacizumab plus carboplatin and paclitaxel (atezo + bev + CP) in clinical practice. The primary endpoint was the 12-month OS rate. Secondary endpoints included OS, progression-free survival, and subgroup analyses, including IMpower-unlike (did not meet the main eligibility criteria of each IMpower trial) and IMpower-like patients. In total, 814 patients were enrolled (atezo + CnP, n = 217; atezo + PP, n = 211; atezo + bev + CP, n = 386). The IMpower-unlike group included patients with Eastern Cooperative Oncology Group performance status ≥ 2, autoimmune disease, or interstitial lung disease. Twelve-month OS rates (95% confidence interval [CI]) were 62.9% (55.8-69.2), 72.1% (65.2-77.9), and 68.3% (63.2-72.9) with atezo + CnP, atezo + PP, and atezo + bev + CP, respectively. OS hazard ratios (95% CI) in the IMpower-unlike vs. -like subgroups were 1.36 (0.91-2.05), 1.08 (0.70-1.68), and 1.49 (1.09-2.06), respectively. No new safety signals were observed. Real-world efficacy and safety for each atezolizumab combination were comparable to those in the relevant IMpower trials.

Electronic patient-reported outcome (ePRO) monitoring for patients undergoing cancer chemotherapy may provide qualified and early detection of adverse events or disease-related symptoms, leading to improved patient care. The aim of this study is to examine whether addition of ePRO monitoring to routine medical care contributes to improved overall survival and quality of life of cancer patients undergoing chemotherapy. Patients with unresectable advanced cancers or metastatic/recurrent solid tumors receiving systemic chemotherapy will be randomized to an ePRO monitoring group and a usual care group. The ePRO group will conduct weekly symptom monitoring using an electronic device after study enrollment until the end of the study. Monitoring results will be returned to medical personnel and used as information for patient care. The primary endpoints are overall survival and health related quality of life. The initial target sample size for the study was 1500 patients. However, due to delays in enrollment, the target was readjusted to 500 patients. Enrollment has been completed, and the study is now in the follow-up phase.

Supplementary Figure S4. OS according to number of whole irAEs Abbreviations: CI, confidence interval; HR, hazard ratio; irAE, immune-related adverse event; NE, not evaluable; NR, not reached; OS, overall survival

Supplementary Table S4. Association between grade 1–2 irAEs and predictors of ICI effect a Mann–Whitney U Test. b Patients who had progression within 4 or 6 weeks after the initiation of treatment were excluded. Abbreviations: CRP, C-reactive protein; ICI, immune checkpoint inhibitor; IHC, immunohistochemical; irAE, immune-related adverse event; NLR, neutrophil-to-lymphocyte ratio; PD-L1, programmed death ligand-1; Q, quartile; SD, standard deviation.

Supplementary Table S5. Association between grade ≥3 irAEs and predictors of ICI effect a Mann–Whitney U Test. b Patients who had progression within 4 or 6 weeks after the initiation of treatment were excluded. Abbreviations: CRP, C-reactive protein; ICI, immune checkpoint inhibitor; IHC, immunohistochemical; irAE, immune-related adverse event; NLR, neutrophil-to-lymphocyte ratio; PD-L1, programmed death ligand-1; Q, quartile; SD, standard deviation.