Ki-67 with Neutrophil-to-lymphocyte and CD4-positive/CD8-positive Ratios Predicts Early Response in Non-small Cell Lung Cancer

Summary

This protocol standardizes early response assessment in NSCLC patients receiving PD-1/PD-L1 inhibitors by integrating baseline NLR, flow-cytometric CD4/CD8 ratio, tumor Ki-67 immunohistochemistry scoring, and RECIST 1.1 evaluation after two cycles, with reproducibility-focused quality-control checkpoints throughout.

Abstract

This study evaluated the combined predictive value of tumor Ki-67 expression status, peripheral blood neutrophil-to-lymphocyte ratio (NLR), and the CD4-positive/CD8-positive T-cell ratio for short-term response assessment in patients with non-small cell lung cancer receiving PD-1/PD-L1 inhibitor therapy. We retrospectively analyzed 102 consecutive patients treated at Binhaiwan Central Hospital between January 2021 and October 2023 who completed two treatment cycles. Responses were assessed using Response Evaluation Criteria in Solid Tumors (RECIST), version 1.1: 13 patients achieved complete response, 26 partial response, 54 stable disease, and 9 progressive disease. For analysis, we defined response as complete response plus partial response and non-response as stable disease plus progressive disease. Univariable screening identified candidate predictors associated with response status, which were entered into multivariable logistic regression adjusted for age, Eastern Cooperative Oncology Group performance status, and smoking history, identifying Ki-67 status, NLR, and the CD4-positive/CD8-positive ratio as independent predictors of short-term response classification. Receiver operating characteristic analysis supported improved discrimination for the combined biomarker model compared with single markers, highlighting a practical, minimally invasive approach for early response stratification in this clinical setting.

Introduction

Accounting for nearly 85% of lung malignancies, non-small cell lung cancer (NSCLC) is the most prevalent histological subtype and predominantly affects middle-aged and older adults in China¹. Its etiology is multifactorial and includes tobacco exposure, long-term contact with industrial dust or pollutants, hereditary susceptibility, and chronic exposure to environmental carcinogens²˒³. Early-stage NSCLC often presents with non-specific respiratory symptoms and can be misattributed to common airway conditions in primary care, contributing to delayed recognition and a high proportion of patients diagnosed at advanced stages⁴˒⁵.

In recent years, immune checkpoint blockade has become a central component of systemic treatment for advanced NSCLC. Programmed death 1 (PD-1) and programmed death ligand 1 (PD-L1) inhibitors have demonstrated durable clinical benefit and improved survival in selected patients⁶. However, response heterogeneity remains substantial, and the financial burden and potential toxicity of prolonged therapy are non-trivial⁷˒⁸. Clinically used predictors such as tumor PD-L1 immunohistochemistry and tumor mutational burden (TMB) provide valuable information, yet both have practical constraints that limit their uniform deployment. PD-L1 assessment can be affected by intratumoral heterogeneity and assay/platform variability, while TMB generally requires sequencing resources, longer turnaround, and higher costs⁹. These constraints make it clinically relevant to develop workflows that use routinely available materials and can be executed with standardized quality control in centers with variable laboratory capacity.

From a practical standpoint, a protocol-driven approach that integrates a small number of accessible biomarkers may support early follow-up planning after treatment initiation. For example, standardized baseline sampling before the first infusion and a defined two-cycle evaluation window can help clinicians identify patients who warrant closer monitoring or earlier reassessment, while maintaining transparency about the steps required to reproduce the measurements across sites. This emphasis on feasibility and reproducibility is particularly relevant for JoVE protocol articles, where the methodological steps—and their quality checkpoints—are central to the contribution.

Ki-67 is a nuclear proliferation marker reflecting tumor growth dynamics and has been discussed in relation to immune contexture. Elevated Ki-67 has been reported to correlate with PD-L1 expression and may be linked to immune escape phenotypes that shape sensitivity to immune checkpoint inhibitors (ICIs)¹⁰. In parallel, the neutrophil-to-lymphocyte ratio (NLR) is a readily accessible index of systemic inflammatory status and has been widely investigated as a prognostic or predictive factor in patients receiving PD-1/PD-L1 inhibitor therapy¹¹. A higher NLR is commonly interpreted as reflecting an immunosuppressive milieu—through relative neutrophilia and/or lymphopenia—and has been associated with poorer response and reduced survival in multiple cohorts¹². The peripheral CD4-positive/CD8-positive T-cell ratio provides an additional window into adaptive immune homeostasis, and the functional balance between helper and cytotoxic T-cell compartments is mechanistically relevant to effective antitumor immunity under PD-1/PD-L1 blockade¹³.

Although Ki-67, NLR, and the CD4-positive/CD8-positive ratio have each shown potential value as individual markers, their integrated use as an operational, minimally invasive workflow for early response assessment in NSCLC remains insufficiently characterized. Therefore, this study examined whether combining tumor Ki-67 expression status with peripheral blood NLR and the CD4-positive/CD8-positive ratio improves short-term response classification after two cycles of PD-1/PD-L1 inhibitor therapy, with the goal of informing early risk stratification and supporting timely clinical reassessment in routine practice.

Protocol

1. Ethics and regulatory compliance

1. Obtain approval from the Ethics Committee of Binhaiwan Central Hospital of Dongguan (Approval No.: 2021045).
2. Ensure all procedures comply with the Declaration of Helsinki.
3. Confirm that written informed consent was obtained prior to treatment and document consent status in the study file.

2. Patient identification and eligibility screening

1. Identify consecutive patients diagnosed with non-small cell lung cancer who initiated programmed cell death protein 1/programmed death-ligand 1 inhibitor therapy between January 2021 and October 2023 at the study center.
2. Confirm pathological diagnosis by reviewing the pathology report.
3. Confirm completion of two treatment cycles and define one cycle as 3 weeks.
4. Apply inclusion criteria.
   1. Confirm age 35–80 years at treatment initiation.
   2. Confirm absence of hepatic metastasis on baseline imaging.
   3. Confirm absence of other primary malignancy.
   4. Confirm documented expected survival time > 12 months in the medical record.
   5. Confirm availability of complete baseline medical records and baseline imaging data.
5. Apply exclusion criteria.
   1. Exclude patients who underwent cardiothoracic surgery within 6 months prior to treatment initiation.
   2. Exclude patients with documented poor compliance compromising follow-up.
   3. Exclude patients with severe psychiatric disorders compromising protocol compliance.
   4. Exclude patients with hematologic or coagulation abnormalities at baseline evaluation.
   5. Exclude patients with severe active infection at baseline.
   6. Exclude patients who received immunotherapy or targeted therapy within 30 days prior to enrollment.
   7. Exclude patients who discontinue treatment during the two-cycle evaluation window due to transfer, financial constraints, or death.
6. Assign a unique study ID to each eligible patient and de-identify all records before analysis.
7. Verify eligibility independently by two investigators and resolve discrepancies by consensus.
8. Record adjudication in a screening log containing study ID, screened item, discrepancy description, adjudicators, and date.
   NOTE: Record baseline conditions likely to shift inflammatory markers (e.g., fever, clinically suspected infection, or systemic corticosteroid exposure) in the screening log to support exclusion decisions and interpretation¹⁴.

3. Baseline clinical data extraction

1. Retrieve baseline variables from the electronic medical record system.
2. Record sex, age, body mass index, education level, tumor location, maximal tumor diameter, tumor–node–metastasis stage, treatment modality, smoking history, and Eastern Cooperative Oncology Group performance status.
3. Export all baseline variables into a de-identified spreadsheet.
4. Lock the dataset for analysis by saving a read-only copy with a timestamp.
5. Verify data accuracy by cross-checking at least 10% of entries against source records.
6. Record corrections in a correction log containing field name, original value, corrected value, reviewer initials, and date.
7. Summarize baseline characteristics by response status and compile the baseline characteristics table for reporting as Table 1.

4. Baseline peripheral blood collection and hematology-derived neutrophil-to-lymphocyte ratio

1. Collect fasting venous blood from an upper-limb vein within 2 h before the first inhibitor administration using EDTA anticoagulant tubes (see Table of Materials).
2. Invert each tube 8–10 times immediately after collection.
3. Run complete blood count testing within 2 h of collection using an automated hematology analyzer (model listed in the Table of Materials).
4. Verify daily internal quality control status for the analyzer before sample testing and record QC pass/fail status, control level, operator, and analyzer model in the study log.
5. Record absolute neutrophil count and absolute lymphocyte count from the analyzer output.
6. Calculate neutrophil-to-lymphocyte ratio using the formula NLR = absolute neutrophil count / absolute lymphocyte count.
7. Flag clotted or hemolyzed samples as invalid in the study log.
8. Repeat blood draw within the same pre-treatment window when clinically feasible and document the reason for any missing or invalid result.
9. Record blood collection time and CBC run time to document the ≤ 2 h pre-analytic window.

5. Flow cytometry for CD4-positive/CD8-positive T-cell ratio

1. Aliquot 50 µL EDTA-anticoagulated whole blood into a 12 × 75 mm flow tube (see Table of Materials).
2. Add 20 µL of a four-color T-cell subset reagent identifying CD3, CD4, CD8, and CD45 using four fluorochromes (see Table of Materials).
3. Vortex for 1–2 s and incubate for 15 min at 20–25 °C protected from light.
4. Add 450 µL of 1× red blood cell lysing solution prepared from a 10× concentrate (see Table of Materials).
5. Mix by gentle inversion and incubate for 10 min at 20–25 °C protected from light.
6. Add 2 mL phosphate-buffered saline (see Table of Materials) and mix by inversion 3–5 times.
7. Centrifuge at 350 × g for 5 min at room temperature.
8. Discard the supernatant without disturbing the pellet.
9. Resuspend the pellet in 300–500 µL phosphate-buffered saline for acquisition.
10. Perform daily cytometer performance checks according to the instrument workflow and record pass/fail status in the run sheet.
11. Set photomultiplier voltages using an unstained tube.
12. Run single-color controls (cells or beads) for each fluorochrome.
13. Calculate compensation within the acquisition software using the single-color controls.
14. Maintain the same voltage settings and compensation matrix for all samples within the same batch and record date, operator, voltage snapshot, and compensation file name.
15. Acquire each sample at a low-to-moderate flow rate and collect at least 10,000 lymphocyte events per sample.
16. Gate lymphocytes using CD45 versus side scatter.
17. Gate singlets using forward scatter area versus forward scatter height.
18. Gate CD3-positive T cells.
19. Quantify CD4-positive and CD8-positive subsets within CD3-positive events.
20. Calculate CD4-positive/CD8-positive ratio using the formula CD4-positive/CD8-positive ratio = (% CD4-positive among CD3-positive) / (% CD8-positive among CD3-positive).
21. Verify separation of CD4-positive and CD8-positive populations on bivariate plots.
22. Recalculate compensation and re-run controls upon observing overlap inconsistent with expected population separation.
23. Save representative plots for each batch (CD45/SSC, singlets, CD3 gate, CD4 vs CD8) and archive them with batch ID and date.
    NOTE: Record reagent lot numbers and run dates in internal run sheets to support batch tracking¹⁵.

6. Tumor Ki-67 immunohistochemistry and binary status assignment

1. Cut formalin-fixed paraffin-embedded tumor tissue sections at 4 µm and mount sections on charged slides (see Table of Materials).
2. Deparaffinize sections in xylene and rehydrate sections through graded ethanol to water.
3. Perform heat-induced epitope retrieval using citrate retrieval buffer, pH 6 (see Table of Materials).
4. Preheat the retrieval buffer to near-boiling temperature before timing.
5. Heat slides in retrieval buffer for 20 min.
6. Cool slides at room temperature for 20 min.
7. Rinse slides with wash buffer to remove residual retrieval solution.
8. Incubate sections with 3% hydrogen peroxide for 10 min at room temperature.
9. Wash slides three times with wash buffer for 1–2 min per wash.
10. Incubate sections with Ki-67 primary antibody (clone MIB-1; ready-to-use formulation; see Table of Materials) for 20 min at room temperature.
11. Wash slides three times with wash buffer for 1–2 min per wash.
12. Apply an HRP-based detection system and develop chromogen with DAB according to the manufacturer instructions (see Table of Materials).
13. Stop chromogen development by rinsing with water once nuclear staining is clearly visible and background remains low.
14. Counterstain sections with hematoxylin, rinse, blue, dehydrate through graded ethanol, clear in xylene, and mount with permanent mounting medium (see Table of Materials).
15. Examine sections under brightfield microscopy at 400×.
16. Select five non-overlapping viable tumor fields and avoid necrosis, crush artifacts, and edge effects.
17. Count at least 100 tumor nuclei per field and record the number of nuclei with unequivocal brown nuclear staining.
18. Count at least 500 tumor nuclei per case across the five fields.
19. Calculate the case-level Ki-67 labeling index using the formula labeling index (%) = (total positive tumor nuclei / total tumor nuclei counted) × 100%.
20. Score staining intensity for each case as 0 (none), 1 (light yellow), 2 (brown-yellow), or 3 (dark brown) using the predominant intensity across counted tumor nuclei.
21. Score positive-cell percentage for each case as 0 (0–5%), 1 (6–25%), 2 (26–50%), 3 (51–75%), or 4 (>75%) using the case-level labeling index.
22. Calculate immunoreactive score using IRS = intensity score × percentage score.
23. Assign Ki-67 expression status for analysis as Ki-67 negative for IRS 0–4 and Ki-67 positive for IRS 5–12, and use this binary status consistently throughout Table 1–4 and modeling outputs.
24. Perform blinded scoring by two pathologists.
25. Blind both pathologists to clinical outcomes and blood biomarker values.
26. Resolve scoring disagreements by joint review and record the final consensus classification and date.
27. Archive representative images for each case with study ID and magnification.
28. Include a known positive control tissue and a negative control (omit primary antibody) in each staining run.
29. Repeat staining upon control failure or upon background staining preventing interpretation.
    NOTE: Archive the scoring worksheet containing five-field counts, labeling index, intensity score, percentage score, IRS, and final negative/positive status for auditability¹⁶.

7. Response evaluation after two cycles

1. Perform baseline imaging before cycle 1 within routine clinical workflow.
2. Repeat imaging after two treatment cycles and define each cycle as 3 weeks.
3. Evaluate treatment response using Response Evaluation Criteria in Solid Tumors, version 1.1.¹⁴
4. Perform independent blinded assessment by two radiologists.
5. Blind both radiologists to biomarker values and modeling covariates.
6. Resolve disagreements by consensus and record adjudication details (reader initials, date, final decision).
7. Classify response category as complete response, partial response, stable disease, or progressive disease according to RECIST 1.1.
8. Assign complete response + partial response to the response group.
9. Assign stable disease + progressive disease to the non-response group.
10. Store baseline and follow-up imaging dates and verify the evaluation window corresponds to two cycles.
11. Archive RECIST measurement sheets with study ID, reader, and date for audit.

8. Statistical analysis

1. Perform statistical analyses using statistical software (see Table of Materials).
2. Present normally distributed continuous variables as mean ± standard deviation and compare groups using an independent-sample t-test.
3. Present categorical variables as counts and percentages and compare groups using the chi-square test or Fisher’s exact test, as appropriate.
4. Code independent variables according to the prespecified coding scheme and compile the coding table as Table 2.
5. Screen predictors using univariable analyses and select variables meeting the prespecified criterion for multivariable modeling.
6. Fit a multivariable logistic regression model using response = 0 and non-response = 1 as the dependent variable.
7. Adjust the multivariable model for age, Eastern Cooperative Oncology Group performance status, and smoking history.
8. Report the final adjusted model estimates as Table 3.
9. Generate receiver operating characteristic curves for single predictors and the combined model.
10. Determine ROC-based thresholds using Youden’s index for NLR and CD4-positive/CD8-positive ratio and apply the derived thresholds consistently in the dichotomized analyses reported in Table 4.
11. Summarize discrimination and classification performance for single predictors and the combined model as Table 4.
12. Perform sensitivity analyses treating neutrophil-to-lymphocyte ratio and CD4-positive/CD8-positive ratio as continuous predictors under the same outcome coding and adjustment set.
13. Report sensitivity analysis results as Supplementary Table S1.
14. Verify consistency between values in the manuscript text and the corresponding tables and figures and archive the final output files and software version used.

Results

Baseline characteristics
Baseline clinical characteristics of the 102 patients with non-small cell lung cancer are summarized in Table 1. The response group (complete response + partial response) included 30 males and 9 females, whereas the non-response group (stable disease + progressive disease) included 50 males and 13 females (p = 0.771). In the response group, 79.49% of patients were younger than 65 years, compared with 66.67% in the non-response group. No statistically signific...

Discussion

Immune checkpoint inhibition has become an essential component of systemic therapy for advanced non-small cell lung cancer (NSCLC), yet clinically meaningful benefit remains concentrated in a subset of patients¹⁵˒¹⁶. This response heterogeneity, together with the cost and potential toxicity of prolonged treatment, makes early response stratification clinically relevant for follow-up planning and timely reassessment¹⁷. In this JoVE protocol, we d...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Automated hematology analyzer	Sysmex Corporation	XN-1000	Complete blood count and NLR calculation
Brightfield microscope	Olympus	BX53	Histological evaluation
Citrate antigen retrieval buffer (pH 6.0)	Abcam	ab93678	Heat-induced antigen retrieval
DAB substrate kit	Vector Laboratories	SK-4100	Chromogenic development
EDTA anticoagulant blood collection tube	BD Biosciences	367525	Peripheral blood collection
Flow cytometer	BD Biosciences	FACSCanto II	T-cell subset analysis
Flow cytometry tubes (12 75 mm)	BD Falcon	352058	Sample acquisition tubes
Formalin-fixed paraffin-embedded tumor tissue blocks	Clinical Pathology Archive	N/A	NSCLC tumor samples
Four-color T-cell subset reagent (CD3/CD4/CD8/CD45)	BD Biosciences	340499	For lymphocyte subset detection
Hematoxylin stain	Sigma-Aldrich	H3136	Nuclear counterstaining
HRP detection kit	Dako (Agilent)	K5007	Secondary antibody detection system
Ki-67 primary antibody (clone MIB-1)	Abcam	ab16667	Immunohistochemical detection
Microtome	Leica Microsystems	RM2235	Cutting 4 m FFPE sections
Phosphate-buffered saline (PBS)	Thermo Fisher Scientific	10010023	Cell washing and resuspension
Positively charged microscope slides	Thermo Fisher Scientific	J1800AMNZ	Tissue section mounting
Red blood cell lysing solution	BD Biosciences	349202	Erythrocyte lysis before flow cytometry
Statistical analysis software	IBM	SPSS v26.0	Data analysis