NCCN-IPI Calculator for DLBCL
NCCN-IPI Calculator for DLBCL
The refined IPI: age in four bands, the LDH ratio in three, and a restricted list of extranodal sites, to a maximum of 8 points. The LDH term is a RATIO to the laboratory’s own upper limit of normal, which is why this index cannot be computed from an absolute LDH.
NCCN-IPI (Zhou 2014)
banded age and LDH ratio → 0–8Age 68; LDH 420 U/L against a laboratory upper limit of 250 U/L; Ann Arbor stage III; no named extranodal site; ECOG 2
The banded items, the LDH ratio and the four strata
Maximum 8 points
Age: ≤ 40 = 0 · >40–60 = 1 · >60–75 = 2 · >75 = 3
LDH / ULN: ≤ 1 = 0 · >1–3 = 1 · >3 = 2
Low 0–1 · low-intermediate 2–3 · high-intermediate 4–5 · high 6–8
- LDH / ULN, and why it is a ratio
- the index was fitted across seven centres with seven different LDH assays, so an absolute threshold would have meant something different at each one. LDH activity depends on the method and on the temperature the reaction is run at, and routine upper limits span roughly 190 to 280 U/L. Normalising to each laboratory’s own limit is the only way the band boundaries transfer. A research group that tried to apply the NCCN-IPI retrospectively to 1,080 patients could not do it at all, for want of the local LDH reference limits
- age in four bands, not two
- the 1993 IPI scored one point for age over 60. The NCCN-IPI scores 1, 2 or 3 across >40–60, >60–75 and >75, so age supplies up to 3 of the 8 points. That single change is most of the improvement in discrimination, and it is also the reason the index has been criticised for over-weighting age in the very old
- a named list, not a count
- bone marrow, central nervous system, liver or gastrointestinal tract, lung. Any one of them scores 1; a patient with three extranodal sites none of which is on the list scores 0. In the validation study read for this page this variable did not independently predict overall survival (p = 0.755), which is worth knowing before weighting it heavily
- a printed disagreement about the top LDH band
- one source read for this page prints the top band as a ratio above 4 rather than above 3, leaving a gap between 3 and 4 in which nothing could score. Two independent sources give above 3, one of them the scoring table of a validation study which states explicitly that the ratio is normalised to each institution’s upper limit. Above 3 is what is implemented
- the cohort and its era
- 1,650 adults with de novo DLBCL diagnosed between 2000 and 2010 at seven NCCN centres, validated in 1,138 patients in British Columbia. This is already a rituximab-era index — unlike FLIPI-1 or the 1993 IPI — but it predates polatuzumab-containing first-line therapy, bispecific antibodies and CAR T-cell therapy in the relapsed setting, so its high-risk group’s 33% five-year survival is a 2000s figure
Worked example
Age 68; LDH 420 U/L against a laboratory upper limit of 250 U/L; Ann Arbor stage III; no named extranodal site; ECOG 2
LDH ratio = 420 / 250 = 1.68 × ULN, which is the >1–3 band and 1 point
Age 68 is the >60–75 band and 2 points
2 + 1 + 1 (stage) + 0 (no named extranodal site) + 1 (ECOG 2) = 5 points
5 sits in the 4–5 high-intermediate band
Now change only the laboratory. At a laboratory whose upper limit of normal is 190 U/L the same measured 420 is a ratio of 2.21 — still the same band, still 1 point. But a measured 700 U/L is a ratio of 2.80 against a 250 limit and 3.68 against a 190 limit: 1 point at one laboratory and 2 at the other, which is the difference between high-intermediate and high risk
And what happens if the ratio is ignored. A reader who entered 420 as though it were the ratio would compute 420 > 3 and score the maximum 2 points for LDH, giving 6 and a high-risk reading. That is the single commonest way this index is miscomputed
Doubling both numbers together — 840 against a 500 limit — leaves the ratio at 1.68 and the score at 5, which is the test that the ratio is behaving as a ratio
Every NCCN-IPI point, and what each one replaced
| Item | Points | 1993 IPI equivalent |
|---|---|---|
| Age 40 or under | 0 | Age 60 or under: 0 |
| Age over 40 to 60 | 1 | Age 60 or under: 0 |
| Age over 60 to 75 | 2 | Age over 60: 1 |
| Age over 75 | 3 | Age over 60: 1 |
| LDH ratio 1 or under | 0 | LDH normal: 0 |
| LDH ratio over 1 to 3 | 1 | LDH above normal: 1 |
| LDH ratio over 3 | 2 | LDH above normal: 1 |
| Ann Arbor stage III or IV | 1 | Same: 1 |
| Bone marrow, CNS, liver or GI tract, or lung involved | 1 | More than one extranodal site: 1 |
| ECOG performance status 2 or worse | 1 | Same: 1 |
The four strata, and the same patients scored by the 1993 index
| Risk group | Score | 5-year OS, NCCN derivation cohort | 5-year OS, an independent validation cohort | Original IPI’s equivalent group |
|---|---|---|---|---|
| Low | 0 – 1 | 96% | 91.1% | 90% |
| Low-intermediate | 2 – 3 | not established here | 79.8% | |
| High-intermediate | 4 – 5 | not established here | 56.8% | |
| High | 6 – 8 | 33% | 31.4% | 54% |
A ratio, not a number, and four age bands instead of two
The NCCN-IPI is the 1993 International Prognostic Index refitted to rituximab-era patients. It keeps the same five variables — age, LDH, stage, extranodal disease and performance status — and changes how three are scored. Age moves from one binary cut at 60 to four bands worth 0 to 3. LDH moves from ‘raised or not’ to three bands of the ratio to normal, worth 0, 1 or 2. The extranodal item stops being a count of sites and becomes a named list of four: bone marrow, central nervous system, liver or gastrointestinal tract, and lung. The maximum goes from 5 points to 8.
The LDH change is the one that decides whether the index can be computed at all. The term is the measured LDH divided by the REPORTING LABORATORY’S OWN upper limit of normal, and it has to be, because LDH is an enzyme activity rather than a concentration: it depends on the method, on the substrate, and on the temperature the reaction is run at, and upper limits in routine use span roughly 190 to 280 U/L. The derivation study pooled seven centres with seven assays and normalised to each. The practical consequence is blunt. One research group that set out to apply the NCCN-IPI to a 1,080-patient cohort reported that it could not do so at all, for want of the local LDH reference limits — and a reader who types an absolute LDH of 420 into a field expecting a ratio scores the maximum 2 points instead of 1 and reads the wrong stratum.
The age banding delivers most of the extra discrimination and is also the index’s most criticised feature: age over 75 scores 3 of the 8 points on its own, so a patient can reach low-intermediate risk on age alone, and modified scorings have been proposed for the very elderly. The extranodal item, by contrast, looks like a refinement and may not be one — in the validation study read for this page it did not independently predict overall survival.
What the index achieved is visible in its own cohort: five-year overall survival of 96% in the low-risk group and 33% in the high-risk group, against 90% and 54% for the 1993 index scored on the same patients. Those figures come from 1,650 adults with de novo DLBCL diagnosed between 2000 and 2010 at seven NCCN centres, validated in 1,138 patients in British Columbia. Every outcome figure here describes the cohort the index was derived in and not the patient in front of you; the spread within one stratum is wider than the gap between strata. They predate polatuzumab-containing first-line therapy, bispecific antibodies and CAR T-cell therapy. Every laboratory threshold here is method- and laboratory-dependent, so the reader’s own laboratory’s reference interval takes precedence. A risk stratum is not a diagnosis and not a plan. This page computes the index, names the stratum and reports what it predicted in the derivation cohort; what follows is a decision for the treating team with the patient.
Frequently asked questions
Why does the NCCN-IPI need the laboratory’s LDH upper limit?
Because the LDH term is a ratio to that limit, not an absolute activity. LDH is measured as an enzyme activity and depends on the method and the assay temperature; routine upper limits span roughly 190 to 280 U/L. A measured 600 U/L is a ratio of 2.4 at one laboratory and 3.2 at another, which is 1 point versus 2. Enter the measured value and the limit from the same report and the unit cancels.
What are the NCCN-IPI age bands?
40 or under scores 0, over 40 to 60 scores 1, over 60 to 75 scores 2, and over 75 scores 3. The 1993 index had a single binary cut at 60 worth one point, so refining age is the main change and supplies much of the improvement in discrimination.
Which extranodal sites count?
Only four: bone marrow, central nervous system, liver or gastrointestinal tract, and lung. Any one of them scores the single point. This replaced the 1993 index’s ‘more than one extranodal site’, so a patient with several extranodal deposits in sites not on the list scores nothing for it.
Is the NCCN-IPI better than the IPI or the R-IPI?
It discriminates better at both ends in the cohorts it has been tested in: in its derivation data the low-risk group’s five-year overall survival was 96% against the IPI’s 90%, and the high-risk group’s 33% against the IPI’s 54%. That is its case. The cost is that it needs the local LDH reference limit, which the IPI and R-IPI do not, and groups without that information cannot compute it.
Is the top LDH band above 3 or above 4?
Above 3. One source read for this page prints above 4, which cannot be right because the band below it ends at 3 and a ratio between 3 and 4 would score nothing. Two independent sources, one of them a validation study’s own scoring table, give above 3.
Related calculators
References
- Zhou Z, et al. An enhanced International Prognostic Index (NCCN-IPI) for patients with diffuse large B-cell lymphoma treated in the rituximab era. Blood. 2014. 1,650 patients diagnosed 2000–2010 at seven NCCN centres, validated in 1,138.
- Oncotarget, article 20988: validation of the NCCN-IPI in DLBCL. Table 1 gives the point values with the top normalised-LDH band as greater than 3; Table 5 gives 5-year overall survival by risk group.
- BMJ Best Practice. Diffuse large B-cell lymphoma — diagnostic criteria: International Prognostic Index, age-adjusted IPI and NCCN-IPI point tables.
- AstraZeneca. Statistical analysis plan for NCT03003520: IPI and NCCN-IPI risk-group ranges and the NCCN-IPI’s 8-point maximum. clinicaltrials.gov.
- Counsell N, et al. Prognostic scores in diffuse large B-cell lymphoma: outcomes by IPI and R-IPI group in 1,080 patients treated with R-CHOP, with the NCCN-IPI left unassessable for want of local LDH reference limits.
Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/