Nottingham Prognostic Index Calculator
Nottingham Prognostic Index Calculator
NPI = 0.2 × size in cm + node stage + grade, each of the last two scored 1 to 3. Derived in 1,629 women at Nottingham City Hospital between 1976 and 1989, so its survival figures belong to that cohort and that era rather than to anyone treated today.
Nottingham Prognostic Index
Size, nodes and grade to an indexInvasive tumour 2.2 cm, node stage 2, grade 2
The index, its coefficient and its two published groupings
Three groups (Haybittle 1982, Galea 1992): good below 3.4 · moderate 3.4 to 5.4 · poor above 5.4
Six groups (Blamey 2007): cut-offs at 2.4, 3.4, 4.4, 5.4 and 6.4
Range: 2.02 points (1 mm, node-negative, grade 1) to 8.0 points (10 cm, four or more nodes, grade 3)
- the 0.2
- the Cox coefficient on size, rounded: Galea’s Table 1 gives 0.17 and the paper prints 0.2. Size is in CENTIMETRES — a 20 mm tumour contributes 0.4 points, and entering it as 20 contributes 4 and moves the patient two prognostic groups
- why node stage and grade carry no coefficient
- they enter with weight 1, so a one-step change in either moves the index a full point while a whole centimetre of tumour moves it 0.2. The categorical terms span 2 to 6 points and size spans 0.02 to 2, so a node-positive grade 3 tumour cannot be in a good prognostic group at any size
- node stage, two codings
- Galea staged by anatomical level — breast only, low axilla or internal mammary alone, apical or both chains — and most later work codes by count. The PLoS validation states it as “1 zero positive nodes, 2 one to three positive nodes, 3 four or more positive nodes”. Where the two differ, say which was used
- the derivation cohort
- Nottingham City Hospital, UK. Fitted on 351 to 387 women treated 1976 to 1981 and validated prospectively in the series that followed; the full published series is 1,629 women treated 1976 to 1989, aged up to 70, with clinically operable tumours under 5 cm
Worked example
Invasive tumour 2.2 cm, node stage 2, grade 2
Size term: 0.2 × 2.2 = 0.44
0.44 + 2 + 2 = 4.44 points
Six-group reading: 4.44 falls in 4.4 to 5.4 → moderate prognostic group II
Three-group reading: 4.44 falls in 3.4 to 5.4 → moderate prognostic group, in which 15-year survival in the original Nottingham series was 42%
Note how little the size is doing. Hold the nodes and grade and grow the tumour from 2.2 cm to 4.2 cm and the index moves only from 4.44 to 4.84 — still the same group
Now change the grade instead. Grade 3 at the same size and node stage gives 5.44, which crosses into the poor prognostic group. One step of grade outweighs five centimetres of tumour
The instrument minimum is 0.2 × 0.1 + 1 + 1 = 2.02 and the maximum is 0.2 × 10 + 3 + 3 = 8.0
The two published groupings, with the cohort each survival figure comes from
| Index | Six-group name | Three-group name | Survival, and whose |
|---|---|---|---|
| Below 2.4 | Excellent | Good (below 3.4) | 97% at 5 years, Charité Berlin, 1,560 women operated 1984–1998 |
| 2.4 to 3.4 | Good | Good (below 3.4) | 80% at 15 years, Nottingham 1976–1989, against 83% in age-matched women |
| 3.4 to 4.4 | Moderate I | Moderate (3.4–5.4) | Combined moderate group: 42% at 15 years, Nottingham; 89% at 5 years, Charité |
| 4.4 to 5.4 | Moderate II | Moderate (3.4–5.4) | As above — the split was made because outcomes within the old moderate group were not homogeneous |
| 5.4 to 6.4 | Poor | Poor (above 5.4) | 13% at 15 years, Nottingham; the combined poor group |
| Above 6.4 | Very poor | Poor (above 5.4) | 48% at 5 years, Charité |
What moves the index, and by how much
| Change | Effect on the index |
|---|---|
| One centimetre of invasive tumour | 0.2 points |
| One step of node stage | 1.0 point |
| One step of histological grade | 1.0 point |
| The whole size range, 1 mm to 10 cm | 1.98 points |
| The whole node and grade range together | 4.0 points |
Three variables, one coefficient, and the unit that catches people
The Nottingham Prognostic Index is one of the oldest prognostic instruments still in routine use, and its durability comes from using only three things a pathology report always contains: the size of the invasive tumour, the extent of nodal involvement and the histological grade. It carries a single coefficient — 0.2 on the size in centimetres — and adds the node stage and grade, each scored 1 to 3, unweighted. The coefficient is the rounded Cox estimate from the derivation paper, which reports 0.17 in its own table and prints 0.2 in the published index.
Because the two categorical terms are unweighted, they dominate. A one-step change in node stage or grade moves the index a full point; an entire centimetre of tumour moves it 0.2. Over operable disease, size contributes about a point in total and nodes and grade four. So a node-positive grade 3 tumour cannot land in a good prognostic group at any size, and the error that most reliably misclassifies a patient is entering the size in millimetres: a 22 mm tumour is 2.2, and entering 22 adds 4.4 instead of 0.44, enough to move someone two groups. The node coding is the one place published practice has diverged: Galea staged by anatomical level — breast only, low axilla or an internal mammary node alone, apical node or both chains — while most later work, including the Charité validation, codes by count. The two agree in most patients and can differ in one with a single apical node.
There are two published groupings and both are in circulation. The original three-group scheme cuts at 3.4 and 5.4, and in the Nottingham series of 1,629 women treated between 1976 and 1989 the good, moderate and poor groups had 15-year survival of 80%, 42% and 13% — the good group’s 80% against 83% in age-matched women from the general population, which is very nearly no excess mortality over fifteen years. The later six-group scheme adds cut-offs at 2.4, 4.4 and 6.4, splitting the old moderate group and separating an excellent group and a very poor one; in the Charité Berlin cohort of 1,560 women operated on between 1984 and 1998, five-year survival ran from 97% to 48%. Those two sets of figures are not comparable with each other and neither is a current prognosis: both cohorts predate modern systemic therapy, and the index takes no account of receptor status, HER2 or treatment. Modelling treatment benefit is what tools such as PREDICT were built for, and NPI+, which adds immunohistochemical classes, is a separate and later instrument rather than a version of this one. A figure from a cohort is not this patient’s outcome and a response category is not a diagnosis: a stratum in which 42 per cent were alive at fifteen years tells you about that stratum, not which 42 per cent. This page computes a published quantity and states the criteria behind it. It renders no dose, no prescription and no treatment decision — that is the treating team’s. Every coefficient, conversion factor and threshold here is attributed to the source it was read in and, where it is a prognostic figure, to its derivation cohort; where the treating protocol differs, the protocol takes precedence.
Frequently asked questions
What is the Nottingham Prognostic Index formula?
NPI = 0.2 multiplied by the invasive tumour size in centimetres, plus the lymph node stage scored 1 to 3, plus the histological grade scored 1 to 3. The 0.2 is the rounded Cox coefficient from Galea and colleagues’ 1992 paper, which gives 0.17 in its own regression table.
Should I enter the tumour size in millimetres or centimetres?
Centimetres. A 22 mm tumour is entered as 2.2; entering 22 adds 4.4 points instead of 0.44, enough to move a patient two prognostic groups, and the result still looks like a plausible index value.
Which prognostic grouping should I use, three groups or six?
Both are published and both are used. The original cuts at 3.4 and 5.4 come from Haybittle 1982 and Galea 1992; the six-group version with cut-offs at 2.4, 3.4, 4.4, 5.4 and 6.4 comes from Blamey and colleagues in 2007. This page shows the six-group band and names the three-group reading alongside it. Say which you are quoting.
How is the lymph node stage defined?
Two codings exist. Galea’s original staged by anatomical level: stage 1 confined to the breast, stage 2 low axillary or internal mammary involvement alone, stage 3 apical node involvement or both chains. Most later work codes by count: 1 for no positive nodes, 2 for one to three, 3 for four or more. They usually agree; record which you used.
Related calculators
References
- Galea MH, Blamey RW, Elston CE, Ellis IO. The Nottingham prognostic index in primary breast cancer. Breast Cancer Res Treat. 1992;22(3):207–19.
- Hartmann-Johnsen OJ, Kåresen R, Schlichting E, Nygård JF. Improving the prognostic ability through better use of standard clinical data — the Nottingham Prognostic Index as an example. PLoS ONE. 2016;11(3):e0149977.
- Classification of patients with breast cancer according to Nottingham Prognostic Index highlights significant differences in immunohistochemical marker expression. World J Surg Oncol. 2014;12:243.
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/