📄 medicine

Development and Validation of a Survival Prediction Model for Lung Squamous Cell Carcinoma

This study developed and validated a simple, clinically applicable survival prediction model for male patients with lung squamous cell carcinoma using the advanced lung cancer inflammation index (ALI) and hemoglobin (Hb) levels, demonstrating its ability to stratify postoperative risk and support individualized treatment strategies.

Original authors: Yoshio Ichihashi, Teruyoshi Amagai, Shinichi Nakatsuka, Kiyoaki Uryuu, Hiromasa Harada, Kaoru Ochi, Ayako Tsunou, Hiroyo Miyata, Kiyoshi Sato, Nobuharu Hanaoka, Takahiro Katsumata

Published 2026-07-10

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Yoshio Ichihashi, Teruyoshi Amagai, Shinichi Nakatsuka, Kiyoaki Uryuu, Hiromasa Harada, Kaoru Ochi, Ayako Tsunou, Hiroyo Miyata, Kiyoshi Sato, Nobuharu Hanaoka, Takahiro Katsumata

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine your body is a bustling city, and lung squamous cell carcinoma (LUSC) is a tricky, unwanted construction project that has taken over a neighborhood. Doctors have successfully demolished the main structure with surgery, but they face a tricky question: "How long will the city remain safe before the trouble returns?"

For a long time, doctors could only guess if a patient would survive for a specific milestone, like "Will you be here in five years?" It's like checking a weather forecast that only says "Rain" or "No Rain" on a specific Tuesday, without telling you how long the storm will actually last. This study, led by researchers like Yoshio Ichihashi, wanted to build a better forecast—one that predicts the actual duration of the storm, measured in days.

The Magic Formula: A Recipe for Time

The researchers looked at 69 male patients who had undergone successful surgery for this specific type of lung cancer. They treated the patients' blood tests and body stats like ingredients in a recipe. They tested many potential ingredients, including inflammation markers and immune system counts, looking for the ones that truly mattered.

They found that two specific ingredients were the stars of the show:

ALI (Advanced Lung Cancer Inflammation Index): Think of this as a "City Health Score." It combines how much fuel the body has (weight), how well the city's supply lines are working (albumin), and how many peacekeepers are on patrol (lymphocytes).
Hb (Hemoglobin): This is the "Oxygen Truck" count, measuring how much oxygen the blood can carry.

By mixing these two ingredients, the team cooked up a simple math formula to predict exactly how many days a patient might survive after surgery:

Survival time (days) = (6.9 × ALI) + (93.4 × Hb) – 198.4

It's like a video game character sheet where your "Health" and "Stamina" stats are plugged into a calculator to tell you your remaining playtime. The study showed that this formula works reasonably well. When they tested it on a new group of patients (an external validation), it correctly ranked patients' survival chances about 70% of the time (a score known as a C-index of 0.70). This suggests the formula is a reliable tool for sorting patients into groups who might need more intense follow-up versus those who might need less.

What Didn't Work: The Dead-End Trail

In their search for the perfect crystal ball, the researchers also looked at a protein called PDPN. Imagine PDPN as a specific type of street sign that some previous studies claimed would warn of future trouble. The team stained tissue samples to see if this sign was present and how bright it glowed.

However, the paper explicitly rules out PDPN as a useful predictor in this specific group. Even though they found that some patients had "strong" signs and others had "weak" or no signs, this difference did not translate to a difference in how long the patients lived. The study suggests that for lung squamous cell carcinoma, this particular street sign is just background noise, not a warning signal. The researchers argue that previous studies might have been looking at a mix of different cancer types, which confused the results, but when they focused strictly on this one type, the sign didn't help predict the future.

How Sure Are We?

The researchers are confident that their new formula is a step forward, but they are careful not to call it a perfect solution. They describe the formula as having "acceptable" discrimination, meaning it's good at telling the difference between a short and a long survival, but it's not a magic 100% crystal ball.

They also admit the study has some limits. It was a "retrospective" look back at past records, which is like reviewing old game logs rather than playing a new game live. The group of 69 patients is relatively small, and because they only looked at men, the formula might not work the same way for everyone. The study suggests that while the formula is a helpful tool for planning, it doesn't prove why these numbers work, and it doesn't guarantee that changing the treatment will change the outcome.

The Takeaway

This study suggests that by simply checking a patient's "City Health Score" (ALI) and "Oxygen Truck" count (Hb), doctors can get a much clearer picture of the future than just guessing at a five-year deadline. It offers a way to personalize care, perhaps helping doctors decide who needs a very close watch and who can relax a bit, all based on a simple calculation of days. But for now, the PDPN sign remains a mystery that didn't solve the puzzle for this specific group of patients.

Technical Summary: Development and Validation of a Survival Prediction Model for Lung Squamous Cell Carcinoma

Problem Statement
Lung cancer remains the leading cause of cancer-related mortality globally. While lung squamous cell carcinoma (LUSC) is a major subtype of non-small-cell lung cancer (NSCLC), postoperative survival varies significantly among patients. Current prognostic methods largely rely on cross-sectional approaches (e.g., 30-day, 5-year, or 10-year survival rates) or binary categorization (alive vs. dead at a specific time point). These methods fail to predict the specific duration of survival for an individual patient. Furthermore, existing biomarker-driven therapies and genetic analyses often lack longitudinal predictive capabilities for individual survival time. There is a clinical need for a simple, accessible tool to predict the exact number of survival days post-surgery to facilitate risk stratification, optimize adjuvant therapy, and personalize follow-up schedules.

Methodology
This retrospective study developed and validated a survival prediction model using readily available clinical biomarkers.

Cohorts: The study included 69 male patients with LUSC who underwent radical surgery. The development cohort consisted of 52 patients from Osaka Medical and Pharmaceutical University Hospital (2007–2015) and 17 from Yao Tokushukai General Hospital (2015–2018). An independent external validation cohort included 35 patients from Osaka Medical and Pharmaceutical University Hospital (up to 2019).
Variables: Data collected included demographics, complete blood counts, biochemistry, and inflammatory indices. Key calculated indices included the Advanced Lung Cancer Inflammation Index (ALI = BMI × Albumin / Neutrophil-to-Lymphocyte Ratio), HALP score, NLR, and SII. Pathological findings (pTNM stage) and Podoplanin (PDPN) immunohistochemical staining were also evaluated.
Statistical Analysis:
- Univariate and multivariate logistic regression were used to identify independent predictors of 5-year survival.
- A Cox proportional hazards regression model was employed to develop a formula for predicting survival time in days.
- Model performance was assessed using Harrell's concordance index (C-index) for discrimination, calibration plots for agreement between predicted and observed outcomes, and bootstrap resampling (1,000 repetitions) for internal validation.
- Decision curve analysis (DCA) was used to evaluate clinical utility.
- Kaplan-Meier analysis was performed to assess the correlation between PDPN staining intensity and survival.

Key Results

Independent Predictors: Multivariate analysis identified the Advanced Lung Cancer Inflammation Index (ALI) and Hemoglobin (Hb) as independent predictors of 5-year survival. Other indices (HALP, NLR, SII) and PDPN staining did not show significant independent predictive value for survival duration in this cohort.
Prediction Formula: The study derived the following linear formula to estimate postoperative survival time in days:
$\text{Survival time (days)} = (6.9 \times \text{ALI}) + (93.4 \times \text{Hb}) - 198.4$
Model Performance:
- Discrimination: The model demonstrated acceptable discrimination with a C-index of 0.70 (95% CI, 0.66–0.74) in the development cohort and 0.65 (95% CI, 0.61–0.69) in the external validation cohort.
- Calibration: Calibration plots showed good agreement between predicted and observed survival probabilities, with a calibration slope of 0.95, indicating minimal overfitting.
- Cutoff Value: ROC analysis identified an optimal ALI cutoff of 48.04 (AUC = 0.690, p = 0.003) for predicting 5-year survival.
PDPN Staining: While PDPN strong positivity showed a trend toward longer survival, the difference was not statistically significant (p = 0.091 for survival time). No significant association was found between PDPN expression and survival outcomes in this specific LUSC cohort.
Clinical Utility: Decision curve analysis indicated a favorable net clinical benefit for using the model.

Significance and Claims
The authors claim this study presents the first survival prediction formula specifically for LUSC that utilizes routinely available clinical variables (ALI and Hb) to estimate individual survival time in days, rather than binary survival status at fixed time points.

The paper asserts that this longitudinal approach offers five specific clinical advantages:

Precision Medicine: Assisting in determining the intensity of postoperative adjuvant therapy.
Therapy Appropriateness: Enabling stratification into low, medium, and high-risk groups to reduce adverse events from under- or overtreatment.
Optimized Follow-up: Allowing for tailored imaging schedules (e.g., intensive monitoring for short predicted survival vs. reduced frequency for long predicted survival) to manage medical costs and resource allocation.
Patient Communication: Providing concrete survival estimates (e.g., median days and confidence intervals) rather than abstract population percentages to support advance care planning and decision-making.
Research Stratification: Serving as a tool for risk stratification in clinical trials and comparative research.

The authors maintain a modest tone regarding limitations, acknowledging the small sample size (69 patients), the retrospective nature of the study, the lack of causal inference, and the need for further research to elucidate the biological mechanisms linking ALI and Hb to LUSC survival. They also note that the model is intended for prognostic estimation and does not imply treatment efficacy.

More like this